From 4262a16ada3f7dc2a3ac7e254d7f67b4b35799b3 Mon Sep 17 00:00:00 2001 From: MacBook Date: Mon, 16 Nov 2026 15:16:54 +0300 Subject: [PATCH] =?UTF-8?q?=D0=A0=D0=B5=D0=B0=D0=BB=D0=B8=D0=B7=D0=BE?= =?UTF-8?q?=D0=B2=D0=B0=D0=BD=D0=B0=20=D0=BB=D0=B0=D0=B1=D0=BE=D1=80=D0=B0?= =?UTF-8?q?=D1=82=D0=BE=D1=80=D0=BD=D0=B0=D1=8F=206=20=D0=BD=D0=B0=20?= =?UTF-8?q?=D0=BE=D1=86=D0=B5=D0=BD=D0=BA=D1=83=204?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .env.example | 9 +- .github/pull_request_template.md | 57 +- Dockerfile | 15 + REPORT.md | 198 +- cmd/messenger/main.go | 370 ++++ compose.yaml | 465 ++++ evidence/clean-start.json | 22 + evidence/correlated-logs.json | 83 + evidence/db-incident-trace.json | 135 ++ evidence/incidents-grade4.json | 63 + evidence/lab6-results.md | 101 + evidence/observability-grade4.json | 29 + evidence/pipeline-grade4.json | 48 + evidence/resilience.json | 156 ++ evidence/trace-example.json | 1918 +++++++++++++++++ evidence/variants-grade4.json | 125 ++ go.mod | 74 + go.sum | 176 ++ infra/Dockerfile.minio | 16 + infra/Dockerfile.worker | 8 + infra/health-aware.sh | 43 + infra/nginx.conf | 35 + infra/otel/README.md | 15 +- infra/otel/alerts.yaml | 23 + infra/otel/collector.yaml | 74 + infra/otel/grafana/dashboards/messenger.json | 669 ++++++ .../provisioning/dashboards/provider.yaml | 11 + .../provisioning/datasources/sources.yaml | 36 + infra/otel/loki.yaml | 30 + infra/otel/prometheus.yaml | 12 + infra/otel/runbook.md | 42 + infra/otel/signals.md | 53 + infra/otel/tempo.yaml | 22 + internal/database/migrate.go | 64 + internal/database/migrations/001_initial.sql | 31 + .../migrations/002_message_idempotency.sql | 13 + .../database/migrations/003_attachments.sql | 30 + .../migrations/004_async_pipeline.sql | 37 + .../database/migrations/005_trace_context.sql | 4 + internal/messenger/attachments.go | 263 +++ internal/messenger/cleanup.go | 118 + internal/messenger/password.go | 44 + internal/messenger/password_test.go | 33 + internal/messenger/server.go | 776 +++++++ internal/objectstore/bootstrap.go | 40 + internal/objectstore/store.go | 108 + internal/pipeline/pipeline.go | 549 +++++ internal/pipeline/pipeline_test.go | 22 + internal/telemetry/dependencies.go | 74 + internal/telemetry/metrics.go | 93 + internal/telemetry/telemetry.go | 205 ++ internal/telemetry/telemetry_test.go | 28 + scripts/init-local.sh | 28 + scripts/restart.sh | 28 + scripts/start-api.sh | 17 + scripts/stop-one.sh | 12 + student-tests/obs_helpers.py | 65 + student-tests/pipeline_helpers.py | 82 + student-tests/verify_incidents.py | 129 ++ student-tests/verify_observability.py | 94 + student-tests/verify_pipeline.py | 138 ++ student-tests/verify_resilience.py | 92 + student-tests/verify_variants.py | 100 + worker/process_image.py | 60 + 64 files changed, 8467 insertions(+), 43 deletions(-) create mode 100644 Dockerfile create mode 100644 cmd/messenger/main.go create mode 100644 compose.yaml create mode 100644 evidence/clean-start.json create mode 100644 evidence/correlated-logs.json create mode 100644 evidence/db-incident-trace.json create mode 100644 evidence/incidents-grade4.json create mode 100644 evidence/lab6-results.md create mode 100644 evidence/observability-grade4.json create mode 100644 evidence/pipeline-grade4.json create mode 100644 evidence/resilience.json create mode 100644 evidence/trace-example.json create mode 100644 evidence/variants-grade4.json create mode 100644 go.mod create mode 100644 go.sum create mode 100644 infra/Dockerfile.minio create mode 100644 infra/Dockerfile.worker create mode 100644 infra/health-aware.sh create mode 100644 infra/nginx.conf create mode 100644 infra/otel/alerts.yaml create mode 100644 infra/otel/collector.yaml create mode 100644 infra/otel/grafana/dashboards/messenger.json create mode 100644 infra/otel/grafana/provisioning/dashboards/provider.yaml create mode 100644 infra/otel/grafana/provisioning/datasources/sources.yaml create mode 100644 infra/otel/loki.yaml create mode 100644 infra/otel/prometheus.yaml create mode 100644 infra/otel/runbook.md create mode 100644 infra/otel/signals.md create mode 100644 infra/otel/tempo.yaml create mode 100644 internal/database/migrate.go create mode 100644 internal/database/migrations/001_initial.sql create mode 100644 internal/database/migrations/002_message_idempotency.sql create mode 100644 internal/database/migrations/003_attachments.sql create mode 100644 internal/database/migrations/004_async_pipeline.sql create mode 100644 internal/database/migrations/005_trace_context.sql create mode 100644 internal/messenger/attachments.go create mode 100644 internal/messenger/cleanup.go create mode 100644 internal/messenger/password.go create mode 100644 internal/messenger/password_test.go create mode 100644 internal/messenger/server.go create mode 100644 internal/objectstore/bootstrap.go create mode 100644 internal/objectstore/store.go create mode 100644 internal/pipeline/pipeline.go create mode 100644 internal/pipeline/pipeline_test.go create mode 100644 internal/telemetry/dependencies.go create mode 100644 internal/telemetry/metrics.go create mode 100644 internal/telemetry/telemetry.go create mode 100644 internal/telemetry/telemetry_test.go create mode 100755 scripts/init-local.sh create mode 100755 scripts/restart.sh create mode 100755 scripts/start-api.sh create mode 100755 scripts/stop-one.sh create mode 100644 student-tests/obs_helpers.py create mode 100644 student-tests/pipeline_helpers.py create mode 100644 student-tests/verify_incidents.py create mode 100644 student-tests/verify_observability.py create mode 100644 student-tests/verify_pipeline.py create mode 100644 student-tests/verify_resilience.py create mode 100644 student-tests/verify_variants.py create mode 100644 worker/process_image.py diff --git a/.env.example b/.env.example index 2b32655..d915c55 100644 --- a/.env.example +++ b/.env.example @@ -7,7 +7,7 @@ SESSION_TTL_SECONDS=3600 # Choose credentials locally. File paths contain no secret values. DATABASE_URL_FILE=/run/secrets/database_url # Optional on 3, required use on 4 in lab2; Redis OR Valkey: -SESSION_STORE=postgres +SESSION_STORE=redis CACHE_URL_FILE=/run/secrets/cache_url API_REPLICAS=2 IDEMPOTENCY_TTL_SECONDS=86400 @@ -19,7 +19,7 @@ S3_SECRET_ACCESS_KEY_FILE=/run/secrets/s3_secret_access_key MAX_UPLOAD_BYTES=10485760 MAX_IMAGE_PIXELS=40000000 BROKER_URL_FILE=/run/secrets/broker_url -WORKER_CONCURRENCY=2 +WORKER_CONCURRENCY=1 JOB_MAX_ATTEMPTS=5 JOB_TIMEOUT_SECONDS=60 STAGING_RETENTION_SECONDS=86400 @@ -28,3 +28,8 @@ OTEL_SERVICE_NAME=messenger-api OTEL_EXPORTER_OTLP_ENDPOINT=http://otel-collector:4318 OTEL_EXPORTER_OTLP_PROTOCOL=http/protobuf OTEL_RESOURCE_ATTRIBUTES=deployment.environment.name=lab + +DB_TIMEOUT_MS=3000 +DB_MAX_CONNS=10 +CLEANUP_GRACE_SECONDS=3600 +GRAFANA_PORT=13000 diff --git a/.github/pull_request_template.md b/.github/pull_request_template.md index 59c9f4e..a77af48 100644 --- a/.github/pull_request_template.md +++ b/.github/pull_request_template.md @@ -1,18 +1,47 @@ ## Лабораторная 6 -Автор / группа: TODO -Целевая оценка (3/4/5): TODO -Трек 5, если заявлен: TODO +Автор / группа: **Меркулова М. А. / БВТ2303** +Целевая оценка: **4**. Трек 5 не заявлен. -Что реализовано и что перенесено: TODO -Команды запуска из чистого clone: TODO -Результаты проверок и ссылка на REPORT.md: TODO -Ограничения / известные ошибки: TODO -Источники и использование ИИ: TODO +### Реализация -- [ ] Весь обязательный минимум текущей и предыдущих работ сохранён с объявленными переходами. -- [ ] Контракт и публичные тесты не ослаблены. -- [ ] make check и make test выполнены; приложен реальный вывод. -- [ ] Сценарии отказов/безопасности своего уровня воспроизведены. -- [ ] Все секреты и приватные ключи находятся вне Git; отчёт не содержит tokens. -- [ ] README задания сохранён, REPORT.md содержит инструкцию именно моего решения. +Перенесена собственная lab5: две Go API-реплики, Nginx, PostgreSQL, Redis sessions, MinIO, Redis Streams/outbox/worker, варианты изображений и обработка отказов. + +Добавлены OTel Go SDK и отдельный Collector; Prometheus, Loki, Tempo и Grafana с persistent volumes и provisioning. HTTP/API, producer/consumer и S3/DB/image spans связаны W3C context через durable job и broker, включая retry. Структурные logs связаны с trace/request IDs; секреты и business payload не экспортируются. Dashboard показывает histogram percentiles, errors/RPS, queue lag/depth, DB pool, worker saturation/heartbeat и CPU/RAM. Есть SLI/SLO, два alert rules, memory limiter/batch/bounded queues и runbook. + +### Запуск + +```sh +sh scripts/init-local.sh +docker compose --env-file .env up -d --build --wait +make check +make test +python3 student-tests/verify_observability.py +python3 student-tests/verify_incidents.py +python3 student-tests/verify_resilience.py +``` + +API по умолчанию 8080, Grafana 13000. При занятом 8080 изменить APP_PORT в .env и передать BASE_URL тестам. Локальная проверка: API 18086, Grafana 13000; Grafana admin password создаётся в secrets и не коммитится. + +### Доказательства + +- [Отчёт](../REPORT.md) +- [Команды и результаты](../evidence/lab6-results.md) +- [Сигналы, trace и logs](../evidence/observability-grade4.json) +- [Инциденты и alert firing/resolved](../evidence/incidents-grade4.json) +- [Dashboard JSON](../infra/otel/grafana/dashboards/messenger.json) +- [Runbook и SLO](../infra/otel/runbook.md) +- [Перезапуски и persistent история](../evidence/resilience.json) + +Worker stop → queue lag firing/resolved; PostgreSQL stop → DB error spans и error-rate firing/resolved. Tempo stop под нагрузкой → бизнес-записи сохранены, queue/RAM ограничены, drops посчитаны. Чистый запуск с новыми secrets и8 пустыми volumes прошёл. + +### Ограничения и источники + +Один Docker-host, нет HA хранилищ; telemetry при отказах может теряться. Sampling ParentBased/AlwaysSample для локального стенда; треки 5 не заявлены. Источники — собственная lab5 и официальные OTel/Grafana/Prometheus/W3C docs; помощь ИИ Codex описана в REPORT.md. + +- [x] Минимум lab5 и бизнес-контракт сохранены. +- [x] Обязательные contracts/tests/README/Makefile не изменены. +- [x] make check, make test, Go tests/vet и собственные проверки выполнены. +- [x] Три сигнала, W3C chain и provisioning проверены на реальных данных. +- [x] Alerts и telemetry backend outage воспроизведены. +- [x] Секреты исключены из Git. diff --git a/Dockerfile b/Dockerfile new file mode 100644 index 0000000..39a7735 --- /dev/null +++ b/Dockerfile @@ -0,0 +1,15 @@ +FROM golang:1.26.8-alpine3.23@sha256:a8fa79c5bd40d880b52bd3b6d7669ecdcfd00e85facdd427d279efb5ddd79cb1 AS build +WORKDIR /src +COPY go.mod go.sum ./ +RUN go mod download +COPY cmd/ ./cmd/ +COPY internal/ ./internal/ +RUN CGO_ENABLED=0 go build -trimpath -ldflags="-s -w" -o /messenger ./cmd/messenger + +FROM scratch +COPY --from=build /messenger /messenger +USER 65532:65532 +ENV APP_HOST=0.0.0.0 APP_PORT=8080 +EXPOSE 8080 +HEALTHCHECK --interval=10s --timeout=3s --start-period=5s --retries=3 CMD ["/messenger", "healthcheck"] +ENTRYPOINT ["/messenger"] diff --git a/REPORT.md b/REPORT.md index b97ae42..8fdc51c 100644 --- a/REPORT.md +++ b/REPORT.md @@ -1,37 +1,191 @@ # Отчёт — лабораторная 6 -- Автор, группа: TODO -- Целевая оценка; выбранный трек 5, если нужен: TODO -- Commit/PR: TODO после публикации -- Использованные источники, библиотеки и помощь ИИ: TODO +- Автор: **Меркулова М. А., БВТ2303**. +- Целевая оценка: **4** — весь минимум и все пункты уровня 4 README. Трек 5 не заявлен. +- Ветка: `solution/BVT2303-Merkulova`, создана от `main` skeleton lab6. +- Основа: собственная реализация lab5. Бизнес-контракт, README задания, COURSE, Makefile, public tests и contracts сохранены. +- Commit/PR заполняется при публикации. Использовалась помощь ИИ Codex при реализации, проверках и документации; источники перечислены ниже. ## Запуск из чистого clone -TODO: версии runtime/образов, prerequisites, команды создания локальных secrets/.env, build, миграций, запуска, ожидания готовности, тестов и остановки. Указать необходимые действия оператора. Не вставлять секреты. Отдельно указать безопасную остановку и команду намеренного удаления учебных данных. +Требования: Docker Desktop с Linux containers, Compose **2.22+** / BuildKit, Python 3.9+, make, OpenSSL и интернет для образов/Go modules/PyPI/исходников MinIO. На Windows — WSL2 с Docker integration; отдельная Windows-проверка не выполнялась. Локальный Go нужен только для Go unit/vet. -## Архитектура и решения +```sh +cd mtusi2026lab6 +sh scripts/init-local.sh +docker compose --env-file .env config --quiet +docker compose --env-file .env up -d --build --wait +make check +make test +``` -TODO: схема процессов/хранилищ, источник истины, транзакционные границы, ограничения. Что перенесено из lab5, что изменено в этой работе. Соответствие переменных из .env.example реальной конфигурации. +`.env` и `secrets/` создаются локально, исключены из Git, не копируются из предыдущей лабы. Миграции выполняет `migrate`, приватный bucket и прикладные права — `s3-init`. Успешные init-контейнеры завершаются с exit 0. Отдельно установленный Redis/БД/MinIO не требуется. Локальные application images собираются из этого репозитория; worker использует build context `application: service:migrate`. + +По умолчанию API опубликован на `http://localhost:8080`, Grafana — `http://localhost:13000`. В проверенном стенде API использует **18086**, поскольку предыдущие лабы занимают 8080/18085. Для этого APP_PORT изменён только в локальном `.env`: + +```sh +make test BASE_URL=http://localhost:18086 +BASE_URL=http://localhost:18086 python3 student-tests/verify_observability.py +BASE_URL=http://localhost:18086 python3 student-tests/verify_incidents.py +``` + +Вход в Grafana: пользователь `admin`, пароль — содержимое локального `secrets/grafana_admin_password`. Пароль не вставляется в отчёт/команды PR. В папке **MTUSI** автоматически появляется dashboard **MTUSI Messenger — lab6**; вручную создавать datasource/panels не нужно. Explore: Prometheus / Loki / Tempo. При изменении GRAFANA_PORT тестам передать `GRAFANA_URL=http://localhost:`. + +`up --wait` проверяет readiness API/хранилищ основного приложения; доступность наблюдаемости дополнительно проверяет `verify_observability.py` через реальные backend API. + +### Проверки + +```sh +make check +make test BASE_URL=http://localhost:18086 +CGO_ENABLED=0 go test ./... +CGO_ENABLED=0 go vet ./... +BASE_URL=http://localhost:18086 python3 student-tests/verify_variants.py +BASE_URL=http://localhost:18086 python3 student-tests/verify_pipeline.py +BASE_URL=http://localhost:18086 python3 student-tests/verify_observability.py +BASE_URL=http://localhost:18086 python3 student-tests/verify_incidents.py +BASE_URL=http://localhost:18086 python3 student-tests/verify_resilience.py +``` + +Тесты изменяют только синтетические данные своего стенда. Сценарии остановки контейнеров запускать последовательно. `verify_observability` создаёт трассу с retry, `verify_incidents` — два обязательных инцидента и отказ Tempo; finally восстанавливает сервисы. `verify_resilience` использует публичные hooks и проверяет StartedAt, данные и сохранённую историю telemetry. + +Остановка без удаления данных: `docker compose --env-file .env down`. Намеренный полный сброс **учебных** данных: `docker compose --env-file .env down --volumes`. + +## Архитектура + +```mermaid +flowchart LR + Client --> Nginx + Nginx --> API1 + Nginx --> API2 + API1 --> PG[(PostgreSQL)] + API2 --> PG + API1 --> R[(Redis sessions)] + API2 --> R + API1 --> S3[(MinIO)] + API2 --> S3 + PG --> Dispatcher + Dispatcher --> Broker[(Redis Streams)] + Broker --> Worker + Worker --> S3 + Worker --> PG + API1 -. OTLP .-> Collector + API2 -. OTLP .-> Collector + Dispatcher -. OTLP .-> Collector + Worker -. OTLP .-> Collector + Collector --> Prometheus + Collector --> Loki + Collector --> Tempo + Grafana --> Prometheus + Grafana --> Loki + Grafana --> Tempo +``` + +- **Prometheus** хранит временные ряды, histogram buckets, recording/alert rules. +- **Loki** хранит структурные логи, принимает native OTLP через `/otlp` (устаревший Loki exporter не используется). +- **Tempo** хранит реальные traces, WAL и blocks. +- **Collector** принимает/пакетирует/экспортирует сигналы, не заменяет долговременное хранилище. +- **Grafana** визуализирует источники; datasource/dashboard provisioning находится в Git. + +Основной API по-прежнему использует PG + Redis sessions + MinIO. Broker/worker и telemetry backend не входят в синхронную readiness API. Принятие job защищено outbox/staging; недоступность telemetry не отменяет business COMMIT. + +### Изменения относительно lab5 + +- `internal/telemetry/`: SDK providers, структурные OTLP logs, безопасные route labels, HTTP/jobs histograms, pool/queue/process/container metrics, DB/Redis tracers. +- `server.go`: server span, X-Trace-Id, request/log correlation; нормализованный route вместо UUID/raw URL. Бизнес-поля ответа не менялись. +- `attachments.go`, migration `005_trace_context.sql`: W3C context и request_id сохраняются вместе с принятием job. +- `pipeline.go`: producer/consumer spans, перенос context через JSON broker, сохранение связи при retry, worker outcome/duration/E2E и busy. +- `store.go`: реальные S3 client spans; SQL acquire/query/COMMIT записываются QueryTracer, без SQL text/args. +- `main.go`: SDK только для API/worker/dispatcher; ограниченный shutdown, register metrics и resource attributes. +- `compose.yaml`, `infra/otel/`: Collector, три backend, Grafana, 4 persistent volumes, provisioning и alerts. +- `scripts/init-local.sh`: локальный пароль Grafana. `restart.sh`: рестарт также пяти telemetry-сервисов. + +### Сквозной контекст + +HTTP span извлекает W3C traceparent/tracestate. Upload сохраняет carrier в той же транзакции, что attachment/jobs/outbox. Dispatcher создаёт **producer child** исходного HTTP span и помещает его carrier в разрешённые поля job. Worker создаёт **consumer child** producer; S3/DB/image.process — дочерние операции consumer. Retry образует новую producer/consumer ветку того же trace. `request_id` хранится в PG для связи лога worker с исходным HTTP ответом, не становится metric label. + +Время до ready — от created_at job в durable acceptance transaction до фиксации ready; включает queue wait и retry. Original/checksum, thumbnail/crop/WebP, ACL, sessions, idempotency, lease и replay lab5 сохранены. + +## Сигналы и dashboard + +Полная схема: [infra/otel/signals.md](infra/otel/signals.md). Исполняемые PromQL/units: [dashboard JSON](infra/otel/grafana/dashboards/messenger.json). + +Dashboard показывает RPS, 5xx/all business ratio, p50/p95/p99 HTTP, worker outcomes/duration, E2E p95 до ready, queued depth/oldest age, pool used/waiting/capacity, worker busy/capacity и heartbeat, container CPU/RAM, DB availability, SLI, alerts и Collector failures/queue. + +Percentiles агрегируются **из суммарных histogram buckets обеих реплик**: + +```promql +histogram_quantile(0.95, + sum by (le) (rate(messenger_http_duration_seconds_bucket{service_name="messenger-api",route!~"/health/.*"}[1m]))) +``` + +Все duration metrics экспортируются в секундах; logs duration_ms явно обозначены. Queue depth — jobs, age — seconds, RAM — bytes, CPU rate — CPU seconds/second (cores). + +Metrics не содержат user_id/chat_id/message_id/attachment_id/request_id/filename/raw URL. В логах и traces допустимы синтетические request/job IDs для корреляции; Authorization, пароли, токены, body и изображение не экспортируются. Это проверено canary-значениями и анализом реальных сохранённых данных. + +## SLI, SLO и alerts + +SLI = доля бизнес HTTP-запросов без 5xx за скользящие **5 минут**; SLO **≥99%**. Health исключены, клиентские 4xx не считаются отказом сервера. При искусственном отказе SLO может нарушаться — это ожидаемое воздействие теста. + +- `MessengerHighErrorRate`: 5xx/all >5% за 1m, `for:10s`. +- `MessengerQueueLag`: oldest queued >10s и depth>0, `for:10s`. +- Scrape 3s, rule evaluation 5s. Доставка наружу не настроена и не требуется: состояния firing/resolved проверяются локально в Prometheus. + +[Alert rules](infra/otel/alerts.yaml) и [runbook](infra/otel/runbook.md) содержат точные запросы, единицы, пороги и шаги диагностики. Alert error ratio — оперативное 1m предупреждение, не burn-rate исследование трека 5. + +## Ограниченная буферизация + +SDK: batch queue 256 spans/log records, batch 64, interval 500 ms, export timeout 2s, без SDK retry; periodic metrics 2s. HTTP handler не ждёт экспорта. Ошибки SDK считаются, payload ошибки не логируется. + +Collector: memory limiter 128 MiB/spike32, hard container limit256 MiB, batch64/max128, по 32 batches в очередях traces/logs, 2 consumers, block_on_overflow=false, retry deadline5s. После исчерпания попыток telemetry теряется, а не блокирует мессенджер. Финальные dropped_items считаются по JSON-логам Collector; failed-send counter включает неуспешные отправки и не приравнивается к числу уникальных трасс. + +## Версии и ресурсы + +| Компонент | Закреплённая версия | CPU / RAM limit | +|---|---|---| +| Go API/dispatcher | Go 1.26.8, builder digest в Dockerfile | API: 0.5 / 1536 MiB; dispatcher 0.25 /128 MiB | +| Worker | Python3.13.7, Pillow12.0.0 + Go binary | 1 /1536 MiB; decoder RLIMIT_AS1GiB | +| PostgreSQL | 17.6-alpine3.22 | 1 /512 MiB | +| Redis sessions / broker | 8.2.1-alpine3.22 | каждый 0.25 /128 MiB | +| MinIO | RELEASE.2025-10-15T17-29-55Z, official source + checksum | 1 /512 MiB | +| Nginx | 1.29.1-alpine3.22 | 0.25 /64 MiB | +| Collector contrib | 0.136.0 | 0.5 /256 MiB | +| Prometheus | v3.7.0 | 0.5 /512 MiB | +| Loki | 3.5.5 | 0.5 /384 MiB | +| Tempo | 2.8.2 | 0.5 /384 MiB | +| Grafana | 12.2.0 | 0.5 /384 MiB | + +SDK traces/metrics1.38.0, logs0.14.0, resource semantic schema1.37.0; собственная схема сигналов v1. Worker concurrency=1. Хранилища metrics/logs/traces имеют retention 24h и отдельные persistent volumes. Grafana state также в volume. Collector очередь in-memory и после его рестарта не восстанавливается; это не бизнес-очередь. + +.env управляет APP_PORT/APP_HOST/APP_ENV, session/idempotency TTL, DB timeout/max connections, worker concurrency/attempts/timeout, staging retention, cleanup grace и GRAFANA_PORT. Внутренний порт API8080, 2 API-реплики, S3 endpoint/region/bucket и OTLP endpoint заданы Compose; ограничения bytes/pixels зафиксированы кодом. Справочные SESSION_STORE/API_REPLICAS/MAX_* из skeleton не переключают реализацию. ## Доказательства критериев -| Критерий README | Команда/сценарий | Наблюдаемый результат | Файл доказательства | -| --- | --- | --- | --- | -| Публичный smoke | make test | TODO | TODO | -| Остальные пункты уровня 3 | TODO | TODO | TODO | -| Все пункты уровня 4, если заявлен | TODO | TODO | TODO | -| Выбранный трек 5, если заявлен | TODO | TODO | TODO | +Актуальные команды/наблюдения — [evidence/lab6-results.md](evidence/lab6-results.md). -TODO: добавить отдельную строку для каждого заявленного критерия. PASS без команды и наблюдения не является доказательством. Проверка инфраструктурного hook требует подтверждения, какие реальные процессы/хранилища перезапущены. +| Критерий | Проверка | Доказательство | +|---|---|---| +| Минимум lab5, бизнес-контракт | make test + verify_variants + verify_pipeline | сохранность оригинала, variants, outbox/retry/replay | +| Три OTLP pipeline и реальные данные | verify_observability.py | observability-grade4.json | +| Server/consumer traces + структурные логи | тот же тест | trace-example.json, correlated-logs.json | +| W3C, producer/consumer, retry, S3/DB | тот же тест | минимум две consumer attempts, проверенные parent IDs | +| Низкая кардинальность и приватность | тот же тест + go test | labels/canary checks | +| Dashboard, percentiles, queue/pool/saturation | provisioning + backend queries | dashboard JSON, реальные metrics | +| SLI/SLO и два alerts | verify_incidents.py | incidents-grade4.json: firing/resolved | +| Два инцидента и диагностика | worker stopped + PostgreSQL stopped | queue/heartbeat, error ratio/DB spans | +| Backend telemetry down, bounded memory/queue | Tempo stopped под нагрузкой | drop counts, memory, business writes | +| Persistent история и core persistence | verify_resilience.py | resilience.json, StartedAt и восстановленная telemetry | -## Эксперименты и отказы +## Ограничения -TODO: гипотеза, hardware/лимиты, версии, объём данных, профиль нагрузки, concurrency/arrival rate, длительность/прогрев, повторы, успешные RPS, ошибки, p50/p95/p99, CPU/RAM. Сохранить raw summary без tokens и объяснить ограничения измерения. +Один Docker-host и одиночные backend/БД/MinIO/broker не обеспечивают HA при потере host/volume. HTTP используется только на localhost с синтетическими данными. Sampling — ParentBased AlwaysSample для локальной демонстрации; стоимость и альтернативы не исследуются. Очереди SDK/Collector ограничены, но telemetry при отказах может теряться. SLO задан и проверяем, это не обещание production SLA. Полный backup/архивирование вне объёма этой лабы. -| Воздействие | Что ожидаем | Что наблюдали | Время восстановления / потеря данных | -| --- | --- | --- | --- | -| TODO | TODO | TODO | TODO | +## Источники -## Ограничения и следующие шаги - -TODO: какие отказы выдерживаются и какие нет; что не реализовано; где учебное упрощение; какое улучшение подтверждено, а какое пока гипотеза. +- Собственная лабораторная 5. +- [OpenTelemetry Go SDK/exporters](https://opentelemetry.io/docs/languages/go/exporters/). +- [W3C Trace Context](https://www.w3.org/TR/trace-context/). +- [Collector configuration](https://opentelemetry.io/docs/collector/configuration/). +- [Loki native OTLP](https://grafana.com/docs/loki/latest/send-data/otel/). +- [Prometheus histograms](https://prometheus.io/docs/practices/histograms/). +- [Grafana provisioning](https://grafana.com/docs/grafana/latest/administration/provisioning/). diff --git a/cmd/messenger/main.go b/cmd/messenger/main.go new file mode 100644 index 0000000..316b160 --- /dev/null +++ b/cmd/messenger/main.go @@ -0,0 +1,370 @@ +package main + +import ( + "context" + "encoding/json" + "errors" + "fmt" + "log" + "net/http" + "os" + "os/signal" + "strconv" + "strings" + "syscall" + "time" + + "github.com/jackc/pgx/v5/pgxpool" + "github.com/minio/minio-go/v7" + "github.com/redis/go-redis/v9" + "mtusi-messenger/internal/database" + "mtusi-messenger/internal/messenger" + "mtusi-messenger/internal/objectstore" + "mtusi-messenger/internal/pipeline" + "mtusi-messenger/internal/telemetry" +) + +func env(name, fallback string) string { + if value := os.Getenv(name); value != "" { + return value + } + return fallback +} +func secret(envName, fileEnv string) (string, error) { + if value := os.Getenv(envName); value != "" { + return value, nil + } + path := os.Getenv(fileEnv) + if path == "" { + return "", fmt.Errorf("set %s or %s", envName, fileEnv) + } + b, err := os.ReadFile(path) + if err != nil { + return "", err + } + value := strings.TrimSpace(string(b)) + if value == "" { + return "", fmt.Errorf("%s is empty", path) + } + return value, nil +} +func positiveInt(name string, fallback int) (int, error) { + raw := env(name, strconv.Itoa(fallback)) + n, err := strconv.Atoi(raw) + if err != nil || n < 1 { + return 0, fmt.Errorf("invalid %s", name) + } + return n, nil +} + +func main() { + if len(os.Args) > 1 && os.Args[1] == "memory" { + values := map[string]int64{} + for _, name := range []string{"memory.current", "memory.peak", "memory.max"} { + data, err := os.ReadFile("/sys/fs/cgroup/" + name) + if err != nil { + log.Fatal("cgroup memory metrics unavailable") + } + n, err := strconv.ParseInt(strings.TrimSpace(string(data)), 10, 64) + if err != nil { + log.Fatal("invalid cgroup metric") + } + values[name] = n + } + _ = json.NewEncoder(os.Stdout).Encode(values) + return + } + port := env("APP_PORT", "8080") + if len(os.Args) > 1 && os.Args[1] == "healthcheck" { + client := &http.Client{Timeout: 2 * time.Second} + resp, err := client.Get("http://127.0.0.1:" + port + "/health/ready") + if err != nil || resp.StatusCode != http.StatusOK { + os.Exit(1) + } + _ = resp.Body.Close() + return + } + host := env("APP_HOST", "0.0.0.0") + if _, err := strconv.Atoi(port); err != nil { + log.Fatal("invalid APP_PORT") + } + ttlSeconds, err := positiveInt("SESSION_TTL_SECONDS", 3600) + if err != nil { + log.Fatal(err) + } + idempotencyTTLSeconds, err := positiveInt("IDEMPOTENCY_TTL_SECONDS", 86400) + if err != nil || idempotencyTTLSeconds < 86400 { + log.Fatal("IDEMPOTENCY_TTL_SECONDS must be at least 86400") + } + dbTimeoutMS, err := positiveInt("DB_TIMEOUT_MS", 3000) + if err != nil { + log.Fatal(err) + } + maxConns, err := positiveInt("DB_MAX_CONNS", 10) + if err != nil { + log.Fatal(err) + } + databaseURL, err := secret("DATABASE_URL", "DATABASE_URL_FILE") + if err != nil { + log.Fatal(err) + } + logger := log.New(os.Stdout, "", log.LstdFlags|log.LUTC) + mode := "api" + if len(os.Args) > 1 { + mode = os.Args[1] + } + if mode == "api" || mode == "worker" || mode == "dispatcher" { + service := "messenger-" + mode + if v := os.Getenv("OTEL_SERVICE_NAME"); v != "" { + service = v + } + initctx, done := context.WithTimeout(context.Background(), 3*time.Second) + shutdown, initErr := telemetry.Start(initctx, service, env("INSTANCE_ID", mode+"-1")) + done() + if initErr != nil { + log.Fatal("invalid telemetry configuration") + } + defer func() { + c, done := context.WithTimeout(context.Background(), 3*time.Second) + defer done() + _ = shutdown(c) + }() + } + + ctx, cancel := context.WithTimeout(context.Background(), 60*time.Second) + defer cancel() + config, err := pgxpool.ParseConfig(databaseURL) + if err != nil { + log.Fatal("invalid database configuration") + } + config.ConnConfig.Tracer = telemetry.DBTracer{} + config.MaxConns = int32(maxConns) + config.ConnConfig.RuntimeParams["application_name"] = env("INSTANCE_ID", "maintenance") + config.MinConns = 1 + config.MaxConnLifetime = 30 * time.Minute + var pool *pgxpool.Pool + for { + pool, err = pgxpool.NewWithConfig(ctx, config) + if err == nil { + err = pool.Ping(ctx) + } + if err == nil { + break + } + if pool != nil { + pool.Close() + } + if ctx.Err() != nil { + log.Fatal("database unavailable during startup") + } + time.Sleep(time.Second) + } + defer pool.Close() + if mode == "api" || mode == "worker" || mode == "dispatcher" { + capacity, _ := positiveInt("WORKER_CONCURRENCY", 1) + if err = telemetry.Monitor(pool, mode, capacity); err != nil { + log.Fatal("invalid telemetry instruments") + } + } + + if len(os.Args) > 1 && os.Args[1] == "migrate" { + if err = database.Migrate(ctx, pool); err != nil { + log.Fatalf("migration failed: %v", err) + } + logger.Print("migrations complete") + return + } + access, err := secret("S3_ACCESS_KEY_ID", "S3_ACCESS_KEY_ID_FILE") + if err != nil { + log.Fatal(err) + } + secretKey, err := secret("S3_SECRET_ACCESS_KEY", "S3_SECRET_ACCESS_KEY_FILE") + if err != nil { + log.Fatal(err) + } + objects, err := objectstore.New(env("S3_ENDPOINT", "http://s3:9000"), env("S3_REGION", "us-east-1"), env("S3_BUCKET", "messenger-attachments"), access, secretKey) + if err != nil { + log.Fatal("invalid S3 configuration") + } + if len(os.Args) > 1 && os.Args[1] == "s3-bootstrap" { + for { + err = objects.Client.MakeBucket(ctx, objects.Bucket, minio.MakeBucketOptions{Region: env("S3_REGION", "us-east-1")}) + if err == nil { + break + } + if exists, e := objects.Client.BucketExists(ctx, objects.Bucket); e == nil && exists { + break + } + if ctx.Err() != nil { + log.Fatal("S3 bootstrap unavailable") + } + time.Sleep(time.Second) + } + if err = objects.Client.SetBucketPolicy(ctx, objects.Bucket, ""); err != nil { + log.Fatal("cannot enforce private bucket") + } + apiAccess, err := secret("S3_API_ACCESS_KEY_ID", "S3_API_ACCESS_KEY_ID_FILE") + if err != nil { + log.Fatal(err) + } + apiSecret, err := secret("S3_API_SECRET_ACCESS_KEY", "S3_API_SECRET_ACCESS_KEY_FILE") + if err != nil { + log.Fatal(err) + } + if err = objectstore.Bootstrap(ctx, env("S3_ENDPOINT", "http://s3:9000"), objects.Bucket, access, secretKey, apiAccess, apiSecret); err != nil { + log.Fatalf("S3 policy bootstrap: %v", err) + } + logger.Print("private MinIO bucket and limited API user ready") + return + } + + if len(os.Args) > 1 && (os.Args[1] == "worker" || os.Args[1] == "dispatcher" || os.Args[1] == "replay" || os.Args[1] == "failed-jobs" || os.Args[1] == "fault") { + + var broker *redis.Client + if os.Args[1] == "worker" || os.Args[1] == "dispatcher" { + brokerURL, e := secret("BROKER_URL", "BROKER_URL_FILE") + if e != nil { + log.Fatal(e) + } + options, e := redis.ParseURL(brokerURL) + if e != nil { + log.Fatal("invalid broker configuration") + } + options.DialTimeout = 2 * time.Second + options.ReadTimeout = 3 * time.Second + options.WriteTimeout = 3 * time.Second + broker = redis.NewClient(options) + defer broker.Close() + } + timeout, e := positiveInt("JOB_TIMEOUT_SECONDS", 60) + if e != nil || timeout > 120 { + log.Fatal("JOB_TIMEOUT_SECONDS must be 1..120") + } + retention, e := positiveInt("STAGING_RETENTION_SECONDS", 86400) + if e != nil { + log.Fatal(e) + } + processor := &pipeline.Pipeline{DB: pool, Broker: broker, Objects: objects, Timeout: time.Duration(timeout) * time.Second, Retention: time.Duration(retention) * time.Second, Logger: logger} + runctx, stop := signal.NotifyContext(context.Background(), syscall.SIGTERM, os.Interrupt) + defer stop() + switch os.Args[1] { + case "dispatcher": + processor.Dispatch(runctx) + case "worker": + concurrency, e := positiveInt("WORKER_CONCURRENCY", 1) + if e != nil { + log.Fatal(e) + } + if e = processor.Worker(runctx, concurrency); e != nil { + log.Fatal(e) + } + default: + if e = processor.Command(ctx, os.Args[1:]); e != nil { + log.Fatal(e) + } + } + return + } + if len(os.Args) > 1 && (os.Args[1] == "cleanup" || os.Args[1] == "cleanup-loop") { + grace, err := positiveInt("CLEANUP_GRACE_SECONDS", 3600) + if err != nil || grace < 120 { + log.Fatal("CLEANUP_GRACE_SECONDS must be at least 120") + } + loopctx, stopCleanup := signal.NotifyContext(context.Background(), syscall.SIGTERM, os.Interrupt) + defer stopCleanup() + for { + cleanupctx, cancelCleanup := context.WithTimeout(loopctx, 60*time.Second) + result, cleanupErr := messenger.Cleanup(cleanupctx, pool, objects, time.Duration(grace)*time.Second) + cancelCleanup() + if cleanupErr != nil { + logger.Print("cleanup failed; retained objects for next sweep") + } else { + _ = json.NewEncoder(os.Stdout).Encode(result) + } + if os.Args[1] == "cleanup" { + if cleanupErr != nil { + os.Exit(1) + } + return + } + select { + case <-loopctx.Done(): + return + case <-time.After(60 * time.Second): + } + } + } + if len(os.Args) > 2 && os.Args[1] == "s3-audit" { + if err = objects.Audit(ctx, os.Args[2]); err != nil { + log.Fatal(err) + } + fmt.Println("PASS: anonymous S3 GET=403; API credentials cannot create buckets") + return + } + if len(os.Args) > 2 && os.Args[1] == "s3-stat" { + info, err := objects.Client.StatObject(ctx, objects.Bucket, os.Args[2], minio.StatObjectOptions{}) + if err != nil { + if minio.ToErrorResponse(err).Code != "NoSuchKey" { + log.Fatal("S3 stat failed") + } + fmt.Println(`{"exists":false}`) + return + } + _ = json.NewEncoder(os.Stdout).Encode(map[string]any{"exists": true, "size_bytes": info.Size}) + return + } + instanceID := env("INSTANCE_ID", "") + if instanceID == "" { + log.Fatal("INSTANCE_ID is required") + } + cacheURL, err := secret("CACHE_URL", "CACHE_URL_FILE") + if err != nil { + log.Fatal(err) + } + redisOptions, err := redis.ParseURL(cacheURL) + if err != nil { + log.Fatal("invalid cache configuration") + } + sessions := redis.NewClient(redisOptions) + sessions.AddHook(telemetry.RedisHook{}) + defer sessions.Close() + for { + if err = sessions.Ping(ctx).Err(); err == nil { + break + } + if ctx.Err() != nil { + log.Fatal("session store unavailable during startup") + } + time.Sleep(time.Second) + } + for { + if err = objects.Ready(ctx); err == nil { + break + } + if ctx.Err() != nil { + log.Fatal("S3 unavailable during startup") + } + time.Sleep(time.Second) + } + dbTimeout := time.Duration(dbTimeoutMS) * time.Millisecond + srv := &http.Server{Addr: host + ":" + port, Handler: messenger.NewHandler(pool, sessions, objects, time.Duration(ttlSeconds)*time.Second, time.Duration(idempotencyTTLSeconds)*time.Second, dbTimeout, instanceID, logger), ReadHeaderTimeout: 5 * time.Second, ReadTimeout: 45 * time.Second, WriteTimeout: 60 * time.Second, IdleTimeout: 30 * time.Second, MaxHeaderBytes: 32 * 1024} + stop := make(chan os.Signal, 1) + signal.Notify(stop, syscall.SIGTERM, os.Interrupt) + done := make(chan struct{}) + go func() { + defer close(done) + <-stop + logger.Print("shutdown started") + shutdownCtx, shutdownCancel := context.WithTimeout(context.Background(), 8*time.Second) + defer shutdownCancel() + if shutdownErr := srv.Shutdown(shutdownCtx); shutdownErr != nil { + logger.Printf("shutdown error: %v", shutdownErr) + } + logger.Print("shutdown complete") + }() + logger.Printf("listening on %s", srv.Addr) + if err = srv.ListenAndServe(); err != nil && !errors.Is(err, http.ErrServerClosed) { + logger.Fatal(err) + } + <-done +} diff --git a/compose.yaml b/compose.yaml new file mode 100644 index 0000000..6adfa55 --- /dev/null +++ b/compose.yaml @@ -0,0 +1,465 @@ +name: mtusi-lab6 + +x-api-common: &api-common + image: mtusi-lab6-api:local + pull_policy: never + environment: &api-environment + APP_HOST: ${APP_HOST:-0.0.0.0} + APP_PORT: 8080 + APP_ENV: ${APP_ENV:-lab} + OTEL_SERVICE_NAME: messenger-api + OTEL_EXPORTER_OTLP_ENDPOINT: http://otel-collector:4318 + OTEL_EXPORTER_OTLP_PROTOCOL: http/protobuf + OTEL_RESOURCE_ATTRIBUTES: deployment.environment.name=lab + SESSION_TTL_SECONDS: ${SESSION_TTL_SECONDS:-3600} + IDEMPOTENCY_TTL_SECONDS: ${IDEMPOTENCY_TTL_SECONDS:-86400} + DB_TIMEOUT_MS: ${DB_TIMEOUT_MS:-3000} + DB_MAX_CONNS: ${DB_MAX_CONNS:-10} + GOMEMLIMIT: 1200MiB + S3_ENDPOINT: http://s3:9000 + S3_REGION: us-east-1 + S3_BUCKET: messenger-attachments + S3_ACCESS_KEY_ID_FILE: /run/secrets/s3_access_key_id + S3_SECRET_ACCESS_KEY_FILE: /run/secrets/s3_secret_access_key + BROKER_URL_FILE: /run/secrets/broker_url + JOB_MAX_ATTEMPTS: ${JOB_MAX_ATTEMPTS:-5} + JOB_TIMEOUT_SECONDS: ${JOB_TIMEOUT_SECONDS:-60} + WORKER_CONCURRENCY: ${WORKER_CONCURRENCY:-1} + STAGING_RETENTION_SECONDS: ${STAGING_RETENTION_SECONDS:-86400} + CLEANUP_GRACE_SECONDS: ${CLEANUP_GRACE_SECONDS:-3600} + DATABASE_URL_FILE: ${DATABASE_URL_FILE:-/run/secrets/database_url} + CACHE_URL_FILE: ${CACHE_URL_FILE:-/run/secrets/cache_url} + secrets: + - database_url + - cache_url + - s3_access_key_id + - s3_secret_access_key + tmpfs: + - /tmp:size=20m,mode=1777 + networks: + - frontend + - backend + healthcheck: + test: ["CMD", "/messenger", "healthcheck"] + interval: 3s + timeout: 2s + retries: 20 + start_period: 3s + restart: unless-stopped + init: true + stop_grace_period: 10s + deploy: + resources: + limits: + cpus: "0.50" + memory: 1536M + +services: + postgres: + image: postgres:17.6-alpine3.22 + environment: + POSTGRES_DB: messenger + POSTGRES_USER: messenger + POSTGRES_PASSWORD_FILE: /run/secrets/postgres_password + secrets: + - postgres_password + volumes: + - postgres_data:/var/lib/postgresql/data + networks: + - backend + healthcheck: + test: ["CMD-SHELL", "pg_isready -U messenger -d messenger"] + interval: 3s + timeout: 3s + retries: 20 + restart: unless-stopped + deploy: + resources: + limits: + cpus: "1.00" + memory: 512M + + redis: + image: redis:8.2.1-alpine3.22 + command: + - /bin/sh + - -ec + - exec redis-server --appendonly yes --appendfsync always --requirepass "$$(cat /run/secrets/redis_password)" + secrets: + - redis_password + volumes: + - redis_data:/data + networks: + - backend + healthcheck: + test: ["CMD-SHELL", "redis-cli --no-auth-warning -a \"$$(cat /run/secrets/redis_password)\" ping | grep PONG"] + interval: 3s + timeout: 3s + retries: 20 + restart: unless-stopped + deploy: + resources: + limits: + cpus: "0.25" + memory: 128M + + migrate: + image: mtusi-lab6-api:local + pull_policy: never + build: + context: . + dockerfile: Dockerfile + healthcheck: + disable: true + command: ["migrate"] + environment: + <<: *api-environment + secrets: + - database_url + networks: + - backend + depends_on: + postgres: + condition: service_healthy + restart: "no" + deploy: + resources: + limits: + cpus: "0.25" + memory: 128M + + api1: + <<: *api-common + environment: + <<: *api-environment + INSTANCE_ID: api-1 + depends_on: + migrate: + condition: service_completed_successfully + redis: + condition: service_healthy + s3-init: + condition: service_completed_successfully + + api2: + <<: *api-common + environment: + <<: *api-environment + INSTANCE_ID: api-2 + depends_on: + migrate: + condition: service_completed_successfully + redis: + condition: service_healthy + s3-init: + condition: service_completed_successfully + + s3: + image: mtusi-lab6-minio:RELEASE.2025-10-15T17-29-55Z + build: + context: . + dockerfile: infra/Dockerfile.minio + command: ["server", "/data", "--address", ":9000", "--console-address", "127.0.0.1:9001"] + environment: + MINIO_ROOT_USER_FILE: /run/secrets/s3_admin_access + MINIO_ROOT_PASSWORD_FILE: /run/secrets/s3_admin_secret + MINIO_BROWSER: "off" + MINIO_UPDATE: "off" + volumes: + - minio_data:/data + secrets: + - s3_admin_access + - s3_admin_secret + networks: + - backend + healthcheck: + test: ["CMD", "wget", "-q", "-T", "1", "-O", "/dev/null", "http://127.0.0.1:9000/minio/health/ready"] + interval: 3s + timeout: 2s + retries: 20 + restart: unless-stopped + deploy: + resources: + limits: + cpus: "1.00" + memory: 512M + + s3-init: + image: mtusi-lab6-api:local + pull_policy: never + healthcheck: + disable: true + command: ["s3-bootstrap"] + environment: + <<: *api-environment + S3_ACCESS_KEY_ID_FILE: /run/secrets/s3_admin_access + S3_SECRET_ACCESS_KEY_FILE: /run/secrets/s3_admin_secret + S3_API_ACCESS_KEY_ID_FILE: /run/secrets/s3_access_key_id + S3_API_SECRET_ACCESS_KEY_FILE: /run/secrets/s3_secret_access_key + secrets: + - database_url + - s3_access_key_id + - s3_secret_access_key + - s3_admin_access + - s3_admin_secret + networks: + - backend + depends_on: + postgres: + condition: service_healthy + s3: + condition: service_healthy + restart: "no" + deploy: + resources: + limits: + cpus: "0.25" + memory: 128M + + cleanup: + image: mtusi-lab6-api:local + pull_policy: never + healthcheck: + disable: true + command: ["cleanup-loop"] + environment: + <<: *api-environment + secrets: + - database_url + - s3_access_key_id + - s3_secret_access_key + networks: + - backend + depends_on: + migrate: + condition: service_completed_successfully + s3-init: + condition: service_completed_successfully + restart: unless-stopped + deploy: + resources: + limits: + cpus: "0.25" + memory: 128M + + broker: + image: redis:8.2.1-alpine3.22 + command: ["/bin/sh", "-ec", 'exec redis-server --appendonly yes --appendfsync always --maxmemory-policy noeviction --requirepass "$$(cat /run/secrets/broker_password)"'] + secrets: [broker_password] + volumes: [broker_data:/data] + networks: [backend] + healthcheck: + test: ["CMD-SHELL", 'redis-cli --no-auth-warning -a "$$(cat /run/secrets/broker_password)" ping | grep PONG'] + interval: 3s + timeout: 2s + retries: 20 + restart: unless-stopped + deploy: + resources: + limits: + cpus: "0.25" + memory: 128M + + dispatcher: + image: mtusi-lab6-api:local + pull_policy: never + command: ["dispatcher"] + healthcheck: + disable: true + environment: + <<: *api-environment + OTEL_SERVICE_NAME: messenger-dispatcher + INSTANCE_ID: dispatcher-1 + secrets: [database_url, s3_access_key_id, s3_secret_access_key, broker_url] + networks: [backend] + depends_on: + migrate: + condition: service_completed_successfully + s3-init: + condition: service_completed_successfully + broker: + condition: service_healthy + restart: unless-stopped + deploy: + resources: + limits: + cpus: "0.25" + memory: 128M + + worker: + image: mtusi-lab6-worker:local + build: + context: . + dockerfile: infra/Dockerfile.worker + additional_contexts: + application: service:migrate + command: ["worker"] + environment: + <<: *api-environment + GOMEMLIMIT: 128MiB + OTEL_SERVICE_NAME: messenger-worker + INSTANCE_ID: worker-1 + secrets: [database_url, s3_access_key_id, s3_secret_access_key, broker_url] + networks: [backend] + tmpfs: ["/tmp:size=128m,mode=1777"] + depends_on: + dispatcher: + condition: service_started + init: true + restart: unless-stopped + stop_grace_period: 5s + deploy: + resources: + limits: + cpus: "1.00" + memory: 1536M + + nginx: + image: nginx:1.29.1-alpine3.22 + command: ["/bin/sh", "/health-aware.sh"] + volumes: + - ./infra/nginx.conf:/etc/nginx/nginx.conf:ro + - ./infra/health-aware.sh:/health-aware.sh:ro + networks: + - frontend + depends_on: + api1: + condition: service_healthy + api2: + condition: service_healthy + ports: + - "127.0.0.1:${APP_PORT:-8080}:8080" + healthcheck: + test: ["CMD-SHELL", "wget -q -O /dev/null http://127.0.0.1:8080/health/ready"] + interval: 3s + timeout: 2s + retries: 20 + restart: unless-stopped + stop_grace_period: 5s + deploy: + resources: + limits: + cpus: "0.25" + memory: 64M + + + otel-collector: + image: otel/opentelemetry-collector-contrib:0.136.0 + command: ["--config=/etc/otelcol/config.yaml"] + volumes: + - ./infra/otel/collector.yaml:/etc/otelcol/config.yaml:ro + networks: [backend] + restart: unless-stopped + deploy: + resources: + limits: + cpus: "0.50" + memory: 256M + + prometheus: + image: prom/prometheus:v3.7.0 + command: ["--config.file=/etc/prometheus/prometheus.yaml", "--storage.tsdb.path=/prometheus", "--storage.tsdb.retention.time=24h"] + volumes: + - ./infra/otel/prometheus.yaml:/etc/prometheus/prometheus.yaml:ro + - ./infra/otel/alerts.yaml:/etc/prometheus/alerts.yaml:ro + - prometheus_data:/prometheus + networks: [backend] + restart: unless-stopped + deploy: + resources: + limits: + cpus: "0.50" + memory: 512M + + loki: + image: grafana/loki:3.5.5 + command: ["-config.file=/etc/loki/config.yaml"] + volumes: + - ./infra/otel/loki.yaml:/etc/loki/config.yaml:ro + - loki_data:/loki + networks: [backend] + restart: unless-stopped + deploy: + resources: + limits: + cpus: "0.50" + memory: 384M + + tempo: + image: grafana/tempo:2.8.2 + command: ["-config.file=/etc/tempo/config.yaml"] + volumes: + - ./infra/otel/tempo.yaml:/etc/tempo/config.yaml:ro + - tempo_data:/var/tempo + networks: [backend] + restart: unless-stopped + deploy: + resources: + limits: + cpus: "0.50" + memory: 384M + + grafana: + image: grafana/grafana:12.2.0 + environment: + GF_SECURITY_ADMIN_USER: admin + GF_SECURITY_ADMIN_PASSWORD__FILE: /run/secrets/grafana_admin_password + GF_USERS_ALLOW_SIGN_UP: "false" + GF_ANALYTICS_REPORTING_ENABLED: "false" + GF_ANALYTICS_CHECK_FOR_UPDATES: "false" + GF_ANALYTICS_CHECK_FOR_PLUGIN_UPDATES: "false" + GF_PLUGINS_PREINSTALL_DISABLED: "true" + GF_AUTH_ANONYMOUS_ENABLED: "false" + secrets: [grafana_admin_password] + volumes: + - ./infra/otel/grafana/provisioning:/etc/grafana/provisioning:ro + - ./infra/otel/grafana/dashboards:/var/lib/grafana/dashboards:ro + - grafana_data:/var/lib/grafana + ports: + - "127.0.0.1:${GRAFANA_PORT:-13000}:3000" + networks: [frontend, backend] + restart: unless-stopped + deploy: + resources: + limits: + cpus: "0.50" + memory: 384M + +secrets: + grafana_admin_password: + file: ./secrets/grafana_admin_password + broker_password: + file: ./secrets/broker_password + broker_url: + file: ./secrets/broker_url + postgres_password: + file: ./secrets/postgres_password + redis_password: + file: ./secrets/redis_password + database_url: + file: ./secrets/database_url + cache_url: + file: ./secrets/cache_url + s3_access_key_id: + file: ./secrets/s3_access_key_id + s3_secret_access_key: + file: ./secrets/s3_secret_access_key + s3_admin_access: + file: ./secrets/s3_admin_access + s3_admin_secret: + file: ./secrets/s3_admin_secret + +volumes: + prometheus_data: + loki_data: + tempo_data: + grafana_data: + broker_data: + postgres_data: + redis_data: + minio_data: + +networks: + frontend: + backend: + internal: true diff --git a/evidence/clean-start.json b/evidence/clean-start.json new file mode 100644 index 0000000..4ced0eb --- /dev/null +++ b/evidence/clean-start.json @@ -0,0 +1,22 @@ +{ + "status": "PASS", + "date": "2026-11-08", + "source": "isolated source copy excluding .git, .env, secrets and virtual environments", + "project": "mtusi-lab6-clean-audit", + "api_port": 18087, + "grafana_port": 13001, + "new_secrets": true, + "fresh_volumes": [ + "postgres_data", + "redis_data", + "minio_data", + "broker_data", + "prometheus_data", + "loki_data", + "tempo_data", + "grafana_data" + ], + "build_cache_used": true, + "public_smoke": "PASS", + "temporary_stack_removed": true +} diff --git a/evidence/correlated-logs.json b/evidence/correlated-logs.json new file mode 100644 index 0000000..d2f5b66 --- /dev/null +++ b/evidence/correlated-logs.json @@ -0,0 +1,83 @@ +[ + { + "attempt": 1, + "event": "job.published", + "job_id": "4985c362-88c5-47f8-8f4e-32e41aebe60a", + "service.name": "messenger-dispatcher", + "severity": "INFO", + "span_id": "65906c44867999d6", + "success": true, + "timestamp": "2026-11-08T15:48:09.340450259Z", + "trace_id": "75692cc8cb7b4ebb920150c60f690298" + }, + { + "attempt": 1, + "event": "job.started", + "job_id": "4985c362-88c5-47f8-8f4e-32e41aebe60a", + "request_id": "c7b498ed-4a65-4858-b5ab-91168efe48b2", + "service.name": "messenger-worker", + "severity": "INFO", + "span_id": "12b2410403dff46d", + "timestamp": "2026-11-08T15:48:09.401979384Z", + "trace_id": "75692cc8cb7b4ebb920150c60f690298" + }, + { + "attempt": 2, + "event": "job.published", + "job_id": "4985c362-88c5-47f8-8f4e-32e41aebe60a", + "service.name": "messenger-dispatcher", + "severity": "INFO", + "span_id": "b3db2d84c0e78904", + "success": true, + "timestamp": "2026-11-08T15:48:14.360894344Z", + "trace_id": "75692cc8cb7b4ebb920150c60f690298" + }, + { + "attempt": 2, + "event": "job.started", + "job_id": "4985c362-88c5-47f8-8f4e-32e41aebe60a", + "request_id": "c7b498ed-4a65-4858-b5ab-91168efe48b2", + "service.name": "messenger-worker", + "severity": "INFO", + "span_id": "2c571d705b0aa552", + "timestamp": "2026-11-08T15:48:14.386684094Z", + "trace_id": "75692cc8cb7b4ebb920150c60f690298" + }, + { + "duration_ms": 52, + "event": "job.completed", + "job_id": "4985c362-88c5-47f8-8f4e-32e41aebe60a", + "outcome": "retry", + "request_id": "c7b498ed-4a65-4858-b5ab-91168efe48b2", + "service.name": "messenger-worker", + "severity": "ERROR", + "span_id": "12b2410403dff46d", + "timestamp": "2026-11-08T15:48:09.45155505Z", + "trace_id": "75692cc8cb7b4ebb920150c60f690298" + }, + { + "duration_ms": 8, + "event": "http.request", + "method": "POST", + "request_id": "c7b498ed-4a65-4858-b5ab-91168efe48b2", + "route": "/api/v1/chats/{chat_id}/attachments", + "service.name": "messenger-api", + "severity": "INFO", + "span_id": "3a0b3d3655700064", + "status": 202, + "timestamp": "2026-11-08T15:48:09.087228092Z", + "trace_id": "75692cc8cb7b4ebb920150c60f690298" + }, + { + "duration_ms": 84, + "event": "job.completed", + "job_id": "4985c362-88c5-47f8-8f4e-32e41aebe60a", + "outcome": "success", + "request_id": "c7b498ed-4a65-4858-b5ab-91168efe48b2", + "service.name": "messenger-worker", + "severity": "INFO", + "span_id": "2c571d705b0aa552", + "timestamp": "2026-11-08T15:48:14.468845719Z", + "trace_id": "75692cc8cb7b4ebb920150c60f690298" + } +] diff --git a/evidence/db-incident-trace.json b/evidence/db-incident-trace.json new file mode 100644 index 0000000..c02564e --- /dev/null +++ b/evidence/db-incident-trace.json @@ -0,0 +1,135 @@ +{ + "batches": [ + { + "resource": { + "attributes": [ + { + "key": "deployment.environment.name", + "value": { + "stringValue": "lab" + } + }, + { + "key": "service.instance.id", + "value": { + "stringValue": "api-2" + } + }, + { + "key": "service.name", + "value": { + "stringValue": "messenger-api" + } + }, + { + "key": "service.version", + "value": { + "stringValue": "lab6-v1" + } + } + ] + }, + "scopeSpans": [ + { + "scope": { + "name": "mtusi-messenger/telemetry-v1" + }, + "spans": [ + { + "traceId": "epY6ExihVHU9tQe8xnXiIw==", + "spanId": "MdlQ/J7piv8=", + "parentSpanId": "bqKvTaWUxEg=", + "flags": 256, + "name": "redis.get", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152654758902803", + "endTimeUnixNano": "1794152654758953261", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "redis" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "get" + } + } + ], + "status": {} + }, + { + "traceId": "epY6ExihVHU9tQe8xnXiIw==", + "spanId": "l0cDzDuR/a4=", + "parentSpanId": "bqKvTaWUxEg=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152654758972844", + "endTimeUnixNano": "1794152654760767552", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "error", + "value": { + "boolValue": true + } + } + ], + "status": { + "message": "dependency unavailable", + "code": "STATUS_CODE_ERROR" + } + }, + { + "traceId": "epY6ExihVHU9tQe8xnXiIw==", + "spanId": "bqKvTaWUxEg=", + "flags": 256, + "name": "POST /api/v1/chats/{chat_id}/messages", + "kind": "SPAN_KIND_SERVER", + "startTimeUnixNano": "1794152654758897136", + "endTimeUnixNano": "1794152654760800886", + "attributes": [ + { + "key": "http.route", + "value": { + "stringValue": "/api/v1/chats/{chat_id}/messages" + } + }, + { + "key": "http.request.method", + "value": { + "stringValue": "POST" + } + }, + { + "key": "request.id", + "value": { + "stringValue": "4cf18920-182a-43f1-9bfc-ab056fed9ce2" + } + }, + { + "key": "http.response.status_code", + "value": { + "intValue": "503" + } + } + ], + "status": { + "message": "HTTP server error", + "code": "STATUS_CODE_ERROR" + } + } + ] + } + ] + } + ] +} diff --git a/evidence/incidents-grade4.json b/evidence/incidents-grade4.json new file mode 100644 index 0000000..2ac9f9d --- /dev/null +++ b/evidence/incidents-grade4.json @@ -0,0 +1,63 @@ +{ + "warmup_load": { + "seconds": 5.001, + "workers": 2, + "successful_requests": 4947, + "failed_requests": 0, + "successful_rps": 989.109, + "latency_ms_successes": { + "p50": 1.937, + "p95": 3.175, + "p99": 4.776 + } + }, + "worker_stopped": { + "firing": true, + "detected_seconds": 25.767, + "depth": 3.0, + "oldest_age_seconds": 24.0, + "worker_busy": 0.0, + "worker_heartbeat_age_seconds": 26.03500008583069, + "resolved": true, + "depth_after": 0.0 + }, + "postgres_stopped": { + "firing": true, + "detected_seconds": 19.539, + "failed_requests": 2584, + "error_ratio": 0.30580225729713156, + "db_available": 0.0, + "pool_waiting_peak": 0.0, + "trace_id": "7a963a1318a154753db507bcc675e223", + "resolved": true, + "db_available_after": 1.0 + }, + "tempo_stopped": { + "workload": { + "seconds": 20.003, + "workers": 2, + "successful_requests": 19732, + "failed_requests": 0, + "successful_rps": 986.464, + "latency_ms_successes": { + "p50": 1.958, + "p95": 3.153, + "p99": 4.537 + } + }, + "collector_before": { + "memory": "71.75MiB / 256MiB", + "cpu": "1.60%" + }, + "collector_during": { + "memory": "68.8MiB / 256MiB", + "cpu": "2.10%" + }, + "failed_span_send_records_delta": 800.0, + "max_queue_batches": 32.0, + "queue_limit_batches": 32, + "dropped_spans_from_collector_logs": 864, + "business_api_preserved": true, + "export_resumed": true + } +} diff --git a/evidence/lab6-results.md b/evidence/lab6-results.md new file mode 100644 index 0000000..b6bf1bf --- /dev/null +++ b/evidence/lab6-results.md @@ -0,0 +1,101 @@ +# Результаты лабораторной 6 — уровень 4 + +**8 ноября 2026 года**, macOS arm64, Docker Desktop Linux VM: Engine 29.2.1, Compose 5.0.2, 10 CPU, RAM VM 8 217 059 328 bytes. Основной API localhost:18086, Grafana localhost:13000. Все данные синтетические; секреты и tokens в evidence не сохраняются. Лимиты сервисов указаны в compose.yaml и REPORT.md. + +## Базовые проверки + +```sh +sh scripts/init-local.sh +docker compose --env-file .env config --quiet +docker compose --env-file .env up -d --build --wait +make check +CGO_ENABLED=0 go test ./... +CGO_ENABLED=0 go vet ./... +docker compose --env-file .env exec -T prometheus promtool check config /etc/prometheus/prometheus.yaml +make test BASE_URL=http://localhost:18086 +``` + +Результат: make check PASS; Go tests messenger/pipeline/telemetry PASS; vet exit0; promtool подтвердил конфигурацию и 4 recording/alert rules; публичный smoke **PASS** при включённой telemetry. + +## Чистый запуск + +Изолированная исходная копия без .env/secrets, новые секреты и **8 пустых volumes**, project `mtusi-lab6-clean-audit`, API 18087/Grafana 13001. `up -d --build --wait` и публичный smoke прошли; BuildKit cache использовался. Проверка выполнена на копии текущих исходников, до публикации ветки в удалённом Git. Временный стенд удалён после проверки, основной сохранён. + +Доказательство: [clean-start.json](clean-start.json). + +## Реальные метрики, логи, трассы и provisioning + +```sh +BASE_URL=http://localhost:18086 python3 student-tests/verify_observability.py +``` + +**PASS**: + +- Контролируемый incoming W3C traceparent сохранён; API ответил тем же X-Trace-Id. +- Trace upload содержит producer и **2 consumer attempts** (временный S3 outage → retry → ready), реальные S3/DB/image.process spans. +- Каждый consumer имеет producer parent span ID; три service resources принадлежат API/dispatcher/worker. +- Структурные события `http.request`, `job.published`, `job.started`, `job.completed` найдены в Loki по одному trace ID; worker log сохранил исходный request_id. +- Сохранённые trace/logs не содержат тестовых пароля, токена, message body и filename. Labels реальных metrics проверены на отсутствие пользовательских/ресурсных IDs. +- Метрики HTTP counter/histogram, worker success/retry/duration, E2E, queue, pool, capacity и память найдены в Prometheus. +- Grafana API подтвердил provisioning dashboard из **23 panels**. Все PromQL targets всех панелей отдельно приняты работающим Prometheus. + +Файлы: [observability-grade4.json](observability-grade4.json), [trace-example.json](trace-example.json), [correlated-logs.json](correlated-logs.json). + +## Два диагностируемых инцидента и отказ telemetry backend + +```sh +BASE_URL=http://localhost:18086 python3 student-tests/verify_incidents.py +``` + +### 1. Worker остановлен + +**Наблюдения из сохранённых сигналов:** queue depth=3, oldest age=24s, busy=0, worker heartbeat age=26.035s. Consumer для новых задач отсутствует, API продолжает сообщения. Это отличает отсутствие worker от медленной обработки внутри него. + +`MessengerQueueLag` перешёл в **firing за 25.767s**. После восстановления те же задания стали ready, depth=0, alert **resolved**. + +### 2. PostgreSQL остановлен + +**Наблюдения:** 5xx ratio **30.58%**, DB availability=0, HTTP trace имеет errored DB child spans. В данном быстром отказе sampled pool waiting peak=0: причина определяется по error spans/availability, а не по предположению, что pool обязательно должен заполниться. + +`MessengerHighErrorRate` перешёл в **firing за 19.539s**; в этом интервале тест получил 2584 ответа 503. После запуска PostgreSQL реальные здоровые запросы восстановились, DB availability=1, alert **resolved**. [Пример DB trace](db-incident-trace.json). + +Диагностика обоих инцидентов опирается на сохранённые Prometheus/Loki/Tempo observations; команды Docker используются для внесения и устранения отказа. SLO: ≥99% бизнес-запросов без 5xx за 5m; контролируемый DB outage намеренно нарушает его. + +### 3. Tempo остановлен + +Во время отказа trace backend выполнялась нагрузка: **20.003s**, 2 closed-loop клиента, **19 732 успешных POST**, **0 HTTP failures**, successful RPS 986.464; успешные latency p50=1.958ms, p95=3.153ms, p99=4.537ms. Перед инцидентами был прогрев 5.001s, 4947 успешных запросов. Это производительность HTTP-подтверждений конкретного локального прогона, не максимальная пропускная способность и не исследование цены telemetry. + +- Максимальная наблюдаемая очередь Collector: **32 batches**, равна настроенному пределу. +- Память Collector: **71.75MiB до / 68.8MiB во время**, hard limit 256MiB; CPU snapshots 1.60% /2.10%. +- Наблюдаемая дельта failed-span-send counter:800 records; JSON-логи финальных drops зафиксировали **864 dropped spans**. Снимки метрик и логов сделаны в разное время; failed attempts не приравниваются к уникальным потерянным traces. +- Потеря signals при bounded retry допускается и посчитана для наблюдаемого Collector drops; бизнес-операции сохранились. +- После запуска Tempo найдена свежая трасса: экспорт восстановился. + +Все три сценария: **PASS**. Подробности: [incidents-grade4.json](incidents-grade4.json). Шаги диагностики: [runbook](../infra/otel/runbook.md). + +## Сохранение возможностей lab5 + +```sh +BASE_URL=http://localhost:18086 python3 student-tests/verify_variants.py +BASE_URL=http://localhost:18086 python3 student-tests/verify_pipeline.py +``` + +Оба **PASS**. Проверены JPEG/WebP, orientation/alpha, bytes/pixels/ACL и checksums; outbox gap, crash после S3 до COMMIT/ack, дубли ready, bounded retry, failed registry, replay, staging cleanup. Expired processing восстановлен за **65.020s**, attempts=2. Пик worker за image suite **704 679 936 bytes** при hard limit1 610 612 736 bytes. + +Файлы: [variants-grade4.json](variants-grade4.json), [pipeline-grade4.json](pipeline-grade4.json). + +## Перезапуск и persistent история + +```sh +BASE_URL=http://localhost:18086 python3 student-tests/verify_resilience.py +``` + +**PASS**. Реально перезапущены **15 core/telemetry сервисов**, StartedAt каждого изменился. Прежние token/data/original сохранились. Ранее сохранённые trace и correlated logs доступны после рестарта Collector/Prometheus/Loki/Tempo/Grafana, метрики снова доступны. + +Отдельно остановлены api1 и api2: подтверждены68/69 сообщений, переходных ошибок0/0, первый успех surviving API0.829s /0.727s, исходный файл сохранён. Обе API-реплики восстановлены. Итоговый `make test` — PASS. + +Доказательство: [resilience.json](resilience.json). + +## Границы проверки + +Проверены уровни 3 и 4 README на одном Docker-host, worker concurrency 1. Полная OpenAPI validation (`make validate`) необязательна локально и в этом прогоне не выполнялась; public smoke проверяет ответы своим schema subset. Показатели latency относятся к успешным запросам закрытого генератора; coordinated omission и влияние клиента не устранены. Отказ host/volume и production HA не заявлены. diff --git a/evidence/observability-grade4.json b/evidence/observability-grade4.json new file mode 100644 index 0000000..eb37f13 --- /dev/null +++ b/evidence/observability-grade4.json @@ -0,0 +1,29 @@ +{ + "status": "PASS", + "trace_id": "75692cc8cb7b4ebb920150c60f690298", + "request_id": "c7b498ed-4a65-4858-b5ab-91168efe48b2", + "consumer_attempt_spans": 2, + "events": [ + "http.request", + "job.completed", + "job.published", + "job.started" + ], + "metrics": { + "http_requests": 40217.0, + "http_histogram": 40217.0, + "jobs_success": 1.0, + "jobs_retry": 1.0, + "ready_histogram": 1.0, + "queue": 0.0, + "pool": 10.0, + "worker_capacity": 1.0, + "memory": 146468864.0 + }, + "dashboard": { + "title": "MTUSI Messenger — lab6", + "panels": 23 + }, + "privacy": "PASS", + "low_cardinality": "PASS" +} diff --git a/evidence/pipeline-grade4.json b/evidence/pipeline-grade4.json new file mode 100644 index 0000000..f0dc4c6 --- /dev/null +++ b/evidence/pipeline-grade4.json @@ -0,0 +1,48 @@ +{ + "stopped_worker_and_restart": { + "status": "PASS", + "before": { + "api1": "2026-11-08T15:37:02.178556088Z", + "api2": "2026-11-08T15:37:02.177388463Z", + "broker": "2026-11-08T15:30:58.77326392Z" + }, + "after": { + "api1": "2026-11-08T15:48:29.295995045Z", + "api2": "2026-11-08T15:48:29.25367792Z", + "broker": "2026-11-08T15:48:29.192814837Z" + } + }, + "commit_publish_gap": "PASS: actual dispatcher exit 86 and automatic restart", + "crash_after_objects": { + "status": "PASS", + "recovery_seconds": 65.02, + "attempts": 2 + }, + "commit_before_ack_and_duplicate": "PASS", + "retry_failed_registry_replay": { + "status": "PASS", + "exhausted_attempts": 2, + "error_code": "retry_exhausted", + "replay_audit_rows": 1, + "ready_replay_rejected": true + }, + "permanent_error": "PASS", + "message_replay_after_attachment_failure": "PASS", + "crash_points": [ + { + "point": "dispatcher_before_publish", + "armed": false, + "hit_at": "2026-11-08T15:48:32.928538+00:00" + }, + { + "point": "worker_after_objects", + "armed": false, + "hit_at": "2026-11-08T15:48:34.168865+00:00" + }, + { + "point": "worker_before_ack", + "armed": false, + "hit_at": "2026-11-08T15:49:40.389531+00:00" + } + ] +} diff --git a/evidence/resilience.json b/evidence/resilience.json new file mode 100644 index 0000000..8238b2d --- /dev/null +++ b/evidence/resilience.json @@ -0,0 +1,156 @@ +{ + "persistence": { + "status": "PASS", + "before": { + "postgres": { + "id": "06797da995bb6cc6fd69a2c251fd6204992b74dbcf630110c59172dde5fb401f", + "started_at": "2026-11-08T15:44:15.618563595Z" + }, + "redis": { + "id": "e2f09b5ae65d8667339960a5649429b6e7abf22d5a20ca9de44846e2394aef6f", + "started_at": "2026-11-08T15:30:58.770536504Z" + }, + "s3": { + "id": "0326a0c248bc89fa1630815b6824af8e03bf57029c88c297bcdf46c2cdc9deb2", + "started_at": "2026-11-08T15:49:47.995548388Z" + }, + "broker": { + "id": "632478d263786461c5e7a9179bdc88c4e910712d58cf30ae2ba0efcc5317a1f3", + "started_at": "2026-11-08T15:48:29.192814837Z" + }, + "api1": { + "id": "b2523eba3a14f02311e1a7dabc2b5e3a0459ce74dc22248559ae919bb33c96bd", + "started_at": "2026-11-08T15:48:29.295995045Z" + }, + "api2": { + "id": "9c3bccf975647eaa0052335faeab360fc199b8aaddc6f1d437009b6a876a1c6f", + "started_at": "2026-11-08T15:48:29.25367792Z" + }, + "dispatcher": { + "id": "d6579f5458efcd35d21d93efe9d64ee27661251275a73137b0cc690f2fc65133", + "started_at": "2026-11-08T15:48:33.042620172Z" + }, + "worker": { + "id": "d5975de80120800bd179bad7284ebf92477094d11b14ae423c159b4d42930f5f", + "started_at": "2026-11-08T15:49:50.782883Z" + }, + "cleanup": { + "id": "aef4f262490e1e616b6c83d8005a1d13286e1b7be5b0389a252315ad66814fdb", + "started_at": "2026-11-08T15:37:02.146939838Z" + }, + "nginx": { + "id": "652643855d5b97cea36ab41ada3ebef618fd894c5a3e9d603759de32b040f65c", + "started_at": "2026-11-08T15:31:08.673401883Z" + }, + "otel-collector": { + "id": "3fcb612ac0c6c3a701d25f66e63b2abfb123ad0390c55f5f2dd795a2f3aa8c2d", + "started_at": "2026-11-08T15:37:11.526894135Z" + }, + "prometheus": { + "id": "a3bad0ed0a03b0d8edc3450830d449ccb6564dfbc3e913674804b82868a9e838", + "started_at": "2026-11-08T15:37:11.547627051Z" + }, + "loki": { + "id": "6713f6f19668716725bd24a7c5e6c8ab6240d2455b2775b854802702f8d24826", + "started_at": "2026-11-08T15:30:58.761601837Z" + }, + "tempo": { + "id": "63377b0fc325bad816738019c191e909751609ca9a5f4ed469896fc6112e0ed2", + "started_at": "2026-11-08T15:45:47.664199012Z" + }, + "grafana": { + "id": "f25430cfedacb0b2a4935bb8a51b63568efffc491fe672b99d3acdd3d02dd6e1", + "started_at": "2026-11-08T15:43:16.869937387Z" + } + }, + "after": { + "postgres": { + "id": "06797da995bb6cc6fd69a2c251fd6204992b74dbcf630110c59172dde5fb401f", + "started_at": "2026-11-08T15:49:56.430728377Z" + }, + "redis": { + "id": "e2f09b5ae65d8667339960a5649429b6e7abf22d5a20ca9de44846e2394aef6f", + "started_at": "2026-11-08T15:49:56.462037627Z" + }, + "s3": { + "id": "0326a0c248bc89fa1630815b6824af8e03bf57029c88c297bcdf46c2cdc9deb2", + "started_at": "2026-11-08T15:49:56.369172627Z" + }, + "broker": { + "id": "632478d263786461c5e7a9179bdc88c4e910712d58cf30ae2ba0efcc5317a1f3", + "started_at": "2026-11-08T15:49:56.495397377Z" + }, + "api1": { + "id": "b2523eba3a14f02311e1a7dabc2b5e3a0459ce74dc22248559ae919bb33c96bd", + "started_at": "2026-11-08T15:50:00.454242921Z" + }, + "api2": { + "id": "9c3bccf975647eaa0052335faeab360fc199b8aaddc6f1d437009b6a876a1c6f", + "started_at": "2026-11-08T15:50:00.410223588Z" + }, + "dispatcher": { + "id": "d6579f5458efcd35d21d93efe9d64ee27661251275a73137b0cc690f2fc65133", + "started_at": "2026-11-08T15:50:00.216710712Z" + }, + "worker": { + "id": "d5975de80120800bd179bad7284ebf92477094d11b14ae423c159b4d42930f5f", + "started_at": "2026-11-08T15:50:01.05319863Z" + }, + "cleanup": { + "id": "aef4f262490e1e616b6c83d8005a1d13286e1b7be5b0389a252315ad66814fdb", + "started_at": "2026-11-08T15:50:00.312024754Z" + }, + "nginx": { + "id": "652643855d5b97cea36ab41ada3ebef618fd894c5a3e9d603759de32b040f65c", + "started_at": "2026-11-08T15:50:06.843809007Z" + }, + "otel-collector": { + "id": "3fcb612ac0c6c3a701d25f66e63b2abfb123ad0390c55f5f2dd795a2f3aa8c2d", + "started_at": "2026-11-08T15:50:06.749767799Z" + }, + "prometheus": { + "id": "a3bad0ed0a03b0d8edc3450830d449ccb6564dfbc3e913674804b82868a9e838", + "started_at": "2026-11-08T15:50:06.911832507Z" + }, + "loki": { + "id": "6713f6f19668716725bd24a7c5e6c8ab6240d2455b2775b854802702f8d24826", + "started_at": "2026-11-08T15:50:07.253593966Z" + }, + "tempo": { + "id": "63377b0fc325bad816738019c191e909751609ca9a5f4ed469896fc6112e0ed2", + "started_at": "2026-11-08T15:50:07.025957216Z" + }, + "grafana": { + "id": "f25430cfedacb0b2a4935bb8a51b63568efffc491fe672b99d3acdd3d02dd6e1", + "started_at": "2026-11-08T15:50:07.067742257Z" + } + } + }, + "failover": [ + { + "status": "PASS", + "stopped_service": "api1", + "confirmed_messages": 68, + "transient_failures": 0, + "first_survivor_seconds": 0.829, + "original_preserved": true + }, + { + "status": "PASS", + "stopped_service": "api2", + "confirmed_messages": 69, + "transient_failures": 0, + "first_survivor_seconds": 0.727, + "original_preserved": true + } + ], + "public_failover": { + "status": "PASS", + "output": "python3 tests/resilience.py --mode failover --action-script \"scripts/stop-one.sh\" --base-url \"http://localhost:18086\" \nExecuting explicit failover hook; hook output suppressed to avoid leaking secrets.\n{\"observed_instances_before\": 2, \"successful_writes\": 131, \"failed_attempts\": 0, \"first_recovery_seconds\": 1.059}\nPASS: failover preserved confirmed data and existing session. Hook scope needs operator evidence." + }, + "telemetry_history_preserved": true, + "metric_history_before_restart": { + "unix_time": 1794152996.911832, + "http_counter_sum": 3161.0 + } +} diff --git a/evidence/trace-example.json b/evidence/trace-example.json new file mode 100644 index 0000000..dd66eb1 --- /dev/null +++ b/evidence/trace-example.json @@ -0,0 +1,1918 @@ +{ + "batches": [ + { + "resource": { + "attributes": [ + { + "key": "deployment.environment.name", + "value": { + "stringValue": "lab" + } + }, + { + "key": "service.instance.id", + "value": { + "stringValue": "api-2" + } + }, + { + "key": "service.name", + "value": { + "stringValue": "messenger-api" + } + }, + { + "key": "service.version", + "value": { + "stringValue": "lab6-v1" + } + } + ] + }, + "scopeSpans": [ + { + "scope": { + "name": "mtusi-messenger/telemetry-v1" + }, + "spans": [ + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "Ogs9NlVwAGQ=", + "traceState": "lab=synthetic", + "parentSpanId": "ASNFZ4mrze8=", + "flags": 768, + "name": "POST /api/v1/chats/{chat_id}/attachments", + "kind": "SPAN_KIND_SERVER", + "startTimeUnixNano": "1794152889079083633", + "endTimeUnixNano": "1794152889087247883", + "attributes": [ + { + "key": "http.route", + "value": { + "stringValue": "/api/v1/chats/{chat_id}/attachments" + } + }, + { + "key": "http.request.method", + "value": { + "stringValue": "POST" + } + }, + { + "key": "request.id", + "value": { + "stringValue": "c7b498ed-4a65-4858-b5ab-91168efe48b2" + } + }, + { + "key": "http.response.status_code", + "value": { + "intValue": "202" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "LVvu1/FmtkY=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "redis.get", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889079096758", + "endTimeUnixNano": "1794152889079194967", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "redis" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "get" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "znhQw0GlfOk=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889079212300", + "endTimeUnixNano": "1794152889079215216", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "f8aGkHp0yqo=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.select", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889079219967", + "endTimeUnixNano": "1794152889079394967", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "SELECT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "akZRtde7cPc=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889080921508", + "endTimeUnixNano": "1794152889080925716", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "EJ+Kunwbdak=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.insert", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889080941008", + "endTimeUnixNano": "1794152889081802883", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "INSERT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "JSkiM3nh+GY=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889081809842", + "endTimeUnixNano": "1794152889081811717", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "8DNzleIUB7A=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.begin", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889081814467", + "endTimeUnixNano": "1794152889081925717", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "BEGIN" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "YCVhK1VHKUo=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.select", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889081929883", + "endTimeUnixNano": "1794152889082255967", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "SELECT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "ezHpQJrzrY0=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.select", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889082259175", + "endTimeUnixNano": "1794152889082516050", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "SELECT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "YOfEItSSy2M=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "s3.put", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889082519133", + "endTimeUnixNano": "1794152889085145050", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "s3" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "ZMVHeXW0qYs=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.insert", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889085161758", + "endTimeUnixNano": "1794152889085860883", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "INSERT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "hB+tP62mLAg=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.delete", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889085865675", + "endTimeUnixNano": "1794152889086033800", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "DELETE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "IrU7B3DrhGQ=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.insert", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889086041925", + "endTimeUnixNano": "1794152889086535133", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "INSERT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "UMtbjNXzX7s=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.insert", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889086546925", + "endTimeUnixNano": "1794152889086932508", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "INSERT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "NktoMJnu5Jc=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 256, + "name": "db.commit", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889086936758", + "endTimeUnixNano": "1794152889087178092", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "COMMIT" + } + } + ], + "status": {} + } + ] + } + ] + }, + { + "resource": { + "attributes": [ + { + "key": "deployment.environment.name", + "value": { + "stringValue": "lab" + } + }, + { + "key": "service.instance.id", + "value": { + "stringValue": "dispatcher-1" + } + }, + { + "key": "service.name", + "value": { + "stringValue": "messenger-dispatcher" + } + }, + { + "key": "service.version", + "value": { + "stringValue": "lab6-v1" + } + } + ] + }, + "scopeSpans": [ + { + "scope": { + "name": "mtusi-messenger/telemetry-v1" + }, + "spans": [ + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "ZZBsRIZ5mdY=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 768, + "name": "attachment.publish", + "kind": "SPAN_KIND_PRODUCER", + "startTimeUnixNano": "1794152889339655634", + "endTimeUnixNano": "1794152889340477093", + "attributes": [ + { + "key": "messaging.system", + "value": { + "stringValue": "redis" + } + }, + { + "key": "messaging.destination.name", + "value": { + "stringValue": "attachment.jobs.v1" + } + }, + { + "key": "job.attempt", + "value": { + "intValue": "1" + } + } + ], + "status": {} + } + ] + } + ] + }, + { + "resource": { + "attributes": [ + { + "key": "deployment.environment.name", + "value": { + "stringValue": "lab" + } + }, + { + "key": "service.instance.id", + "value": { + "stringValue": "worker-1" + } + }, + { + "key": "service.name", + "value": { + "stringValue": "messenger-worker" + } + }, + { + "key": "service.version", + "value": { + "stringValue": "lab6-v1" + } + } + ] + }, + "scopeSpans": [ + { + "scope": { + "name": "mtusi-messenger/telemetry-v1" + }, + "spans": [ + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "ErJBBAPf9G0=", + "traceState": "lab=synthetic", + "parentSpanId": "ZZBsRIZ5mdY=", + "flags": 768, + "name": "attachment.consume", + "kind": "SPAN_KIND_CONSUMER", + "startTimeUnixNano": "1794152889399092092", + "endTimeUnixNano": "1794152889451575467", + "attributes": [ + { + "key": "messaging.system", + "value": { + "stringValue": "redis" + } + }, + { + "key": "messaging.destination.name", + "value": { + "stringValue": "attachment.jobs.v1" + } + }, + { + "key": "job.id", + "value": { + "stringValue": "4985c362-88c5-47f8-8f4e-32e41aebe60a" + } + }, + { + "key": "job.attempt", + "value": { + "intValue": "1" + } + }, + { + "key": "job.outcome", + "value": { + "stringValue": "retry" + } + } + ], + "status": { + "message": "retry", + "code": "STATUS_CODE_ERROR" + } + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "8QnZ9JIaq4Y=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889399103592", + "endTimeUnixNano": "1794152889399106467", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "9qYsw4oYa5g=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.select", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889399117592", + "endTimeUnixNano": "1794152889399617759", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "SELECT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "o//B9UUFX6I=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.begin", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889399622217", + "endTimeUnixNano": "1794152889399659384", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "BEGIN" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "b0OmW7pT9rw=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.select", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889399663425", + "endTimeUnixNano": "1794152889400710842", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "SELECT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "ID7vNM/3FfY=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889400720509", + "endTimeUnixNano": "1794152889401246175", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "Q0bwc+WL1/8=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889401250467", + "endTimeUnixNano": "1794152889401636800", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "US12D+et36s=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.commit", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889401639884", + "endTimeUnixNano": "1794152889401977968", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "COMMIT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "JexrXu211Bc=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889402011592", + "endTimeUnixNano": "1794152889402013259", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "4XgcAkdLLgo=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889402015342", + "endTimeUnixNano": "1794152889402577467", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "Sr4VeShyDb0=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "s3.get", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889402590842", + "endTimeUnixNano": "1794152889448936758", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "s3" + } + }, + { + "key": "error", + "value": { + "boolValue": true + } + } + ], + "status": { + "message": "dependency unavailable", + "code": "STATUS_CODE_ERROR" + } + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "KJtzCpBcaQI=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889448968092", + "endTimeUnixNano": "1794152889448970800", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "AUif6/q4Yus=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889448974842", + "endTimeUnixNano": "1794152889449190633", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "E2jq0upUoaM=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.begin", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889449199634", + "endTimeUnixNano": "1794152889449273176", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "BEGIN" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "j8tb1OaplJ0=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.select", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889449276217", + "endTimeUnixNano": "1794152889449499633", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "SELECT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "Wtofas+h2+M=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889449503009", + "endTimeUnixNano": "1794152889449790384", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "Chg01KPmS0g=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889449793509", + "endTimeUnixNano": "1794152889449996509", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "SUwCJGPUHdY=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.commit", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889450000092", + "endTimeUnixNano": "1794152889450271509", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "COMMIT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "6vaLSWXl1Yk=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889450275717", + "endTimeUnixNano": "1794152889450276717", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "jhlIQJrZsyk=", + "traceState": "lab=synthetic", + "parentSpanId": "ErJBBAPf9G0=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152889450280175", + "endTimeUnixNano": "1794152889450382092", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + } + ] + } + ] + }, + { + "resource": { + "attributes": [ + { + "key": "deployment.environment.name", + "value": { + "stringValue": "lab" + } + }, + { + "key": "service.instance.id", + "value": { + "stringValue": "dispatcher-1" + } + }, + { + "key": "service.name", + "value": { + "stringValue": "messenger-dispatcher" + } + }, + { + "key": "service.version", + "value": { + "stringValue": "lab6-v1" + } + } + ] + }, + "scopeSpans": [ + { + "scope": { + "name": "mtusi-messenger/telemetry-v1" + }, + "spans": [ + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "s9sthMDniQQ=", + "traceState": "lab=synthetic", + "parentSpanId": "Ogs9NlVwAGQ=", + "flags": 768, + "name": "attachment.publish", + "kind": "SPAN_KIND_PRODUCER", + "startTimeUnixNano": "1794152894356629344", + "endTimeUnixNano": "1794152894360999094", + "attributes": [ + { + "key": "messaging.system", + "value": { + "stringValue": "redis" + } + }, + { + "key": "messaging.destination.name", + "value": { + "stringValue": "attachment.jobs.v1" + } + }, + { + "key": "job.attempt", + "value": { + "intValue": "2" + } + } + ], + "status": {} + } + ] + } + ] + }, + { + "resource": { + "attributes": [ + { + "key": "deployment.environment.name", + "value": { + "stringValue": "lab" + } + }, + { + "key": "service.instance.id", + "value": { + "stringValue": "worker-1" + } + }, + { + "key": "service.name", + "value": { + "stringValue": "messenger-worker" + } + }, + { + "key": "service.version", + "value": { + "stringValue": "lab6-v1" + } + } + ] + }, + "scopeSpans": [ + { + "scope": { + "name": "mtusi-messenger/telemetry-v1" + }, + "spans": [ + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "LFcdcFsKpVI=", + "traceState": "lab=synthetic", + "parentSpanId": "s9sthMDniQQ=", + "flags": 768, + "name": "attachment.consume", + "kind": "SPAN_KIND_CONSUMER", + "startTimeUnixNano": "1794152894384762553", + "endTimeUnixNano": "1794152894468861887", + "attributes": [ + { + "key": "messaging.system", + "value": { + "stringValue": "redis" + } + }, + { + "key": "messaging.destination.name", + "value": { + "stringValue": "attachment.jobs.v1" + } + }, + { + "key": "job.id", + "value": { + "stringValue": "4985c362-88c5-47f8-8f4e-32e41aebe60a" + } + }, + { + "key": "job.attempt", + "value": { + "intValue": "2" + } + }, + { + "key": "job.outcome", + "value": { + "stringValue": "success" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "TWJjNYSJw8M=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "image.process", + "kind": "SPAN_KIND_INTERNAL", + "startTimeUnixNano": "1794152894393726844", + "endTimeUnixNano": "1794152894461340802", + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "sgX6RB7/cZI=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894461380886", + "endTimeUnixNano": "1794152894461384344", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "VWXdCGj3Sec=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.select", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894461387719", + "endTimeUnixNano": "1794152894461662011", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "SELECT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "l1Z7QauNHW4=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "s3.put", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894461685469", + "endTimeUnixNano": "1794152894464154094", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "s3" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "CDRT/kXBlvc=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "s3.put", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894464162261", + "endTimeUnixNano": "1794152894465363178", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "s3" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "1BTQsYbRY5Y=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "s3.put", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894465371261", + "endTimeUnixNano": "1794152894466330511", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "s3" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "snHIRaKlq8U=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "s3.put", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894466341094", + "endTimeUnixNano": "1794152894467270386", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "s3" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "tfTVBV5SkEo=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894467298053", + "endTimeUnixNano": "1794152894467299178", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "8WssAMC/2DI=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894467301803", + "endTimeUnixNano": "1794152894467403969", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "jxLWv+PwCY4=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.begin", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894467407636", + "endTimeUnixNano": "1794152894467452427", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "BEGIN" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "jRuTsXWXIVg=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.select", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894467454636", + "endTimeUnixNano": "1794152894467536427", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "SELECT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "Hpck/fZgclc=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894467544094", + "endTimeUnixNano": "1794152894467719719", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "OARkD11KiKk=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894467721928", + "endTimeUnixNano": "1794152894467834053", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "JyQdwMd8npA=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.commit", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894467835636", + "endTimeUnixNano": "1794152894467954136", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "COMMIT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "ml8ICPbqzTc=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894467963719", + "endTimeUnixNano": "1794152894467964802", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "e9cncSnRRwg=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894467966344", + "endTimeUnixNano": "1794152894468028677", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + } + ] + } + ] + }, + { + "resource": { + "attributes": [ + { + "key": "deployment.environment.name", + "value": { + "stringValue": "lab" + } + }, + { + "key": "service.instance.id", + "value": { + "stringValue": "worker-1" + } + }, + { + "key": "service.name", + "value": { + "stringValue": "messenger-worker" + } + }, + { + "key": "service.version", + "value": { + "stringValue": "lab6-v1" + } + } + ] + }, + "scopeSpans": [ + { + "scope": { + "name": "mtusi-messenger/telemetry-v1" + }, + "spans": [ + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "pajIOC6tHsQ=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894384787053", + "endTimeUnixNano": "1794152894385055970", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "2lAStMbUZzM=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.select", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894385079303", + "endTimeUnixNano": "1794152894385279845", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "SELECT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "zI/zvIey9vo=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.begin", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894385286803", + "endTimeUnixNano": "1794152894385359762", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "BEGIN" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "izVuw25iOig=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.select", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894385371053", + "endTimeUnixNano": "1794152894385712887", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "SELECT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "MnIHDYi6w+0=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894385736803", + "endTimeUnixNano": "1794152894386078428", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "DlkhSq1YJQ4=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894386096469", + "endTimeUnixNano": "1794152894386378011", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "b1V5JqJ93zk=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.commit", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894386383136", + "endTimeUnixNano": "1794152894386677636", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "COMMIT" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "HdG/Z1Yoem4=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894386761386", + "endTimeUnixNano": "1794152894386960636", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "fueadYYQ9UE=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.update", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894386977803", + "endTimeUnixNano": "1794152894387206511", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "UPDATE" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "mZzf8tklICI=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "s3.get", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894387259178", + "endTimeUnixNano": "1794152894389932261", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "s3" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "fJQZxyNiwIc=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.pool.acquire", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894391776969", + "endTimeUnixNano": "1794152894391781136", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + } + ], + "status": {} + }, + { + "traceId": "dWksyMt7TruSAVDGD2kCmA==", + "spanId": "O1B64kMcnA0=", + "traceState": "lab=synthetic", + "parentSpanId": "LFcdcFsKpVI=", + "flags": 256, + "name": "db.select", + "kind": "SPAN_KIND_CLIENT", + "startTimeUnixNano": "1794152894391877136", + "endTimeUnixNano": "1794152894393534177", + "attributes": [ + { + "key": "db.system.name", + "value": { + "stringValue": "postgresql" + } + }, + { + "key": "db.operation.name", + "value": { + "stringValue": "SELECT" + } + } + ], + "status": {} + } + ] + } + ] + } + ] +} diff --git a/evidence/variants-grade4.json b/evidence/variants-grade4.json new file mode 100644 index 0000000..036a705 --- /dev/null +++ b/evidence/variants-grade4.json @@ -0,0 +1,125 @@ +{ + "orientation": { + "status": "PASS", + "variants": { + "thumbnail": { + "content_type": "image/jpeg", + "height": 256, + "name": "thumbnail", + "sha256": "335b4953e981a975642374e12196a4701e353972f223c1a7e850ac5998588f1a", + "size_bytes": 1270, + "width": 160 + }, + "crop": { + "content_type": "image/jpeg", + "height": 128, + "name": "crop", + "sha256": "b7f4884df301163117ebb3189f6141a9e376cc6d94804e2f56ce38e0dc3dd047", + "size_bytes": 886, + "width": 128 + }, + "optimized": { + "content_type": "image/webp", + "height": 320, + "name": "optimized", + "sha256": "577722d04127a774de08f674717fa03e3058124395f9ee31a949e6f508e86b43", + "size_bytes": 210, + "width": 200 + } + } + }, + "alpha": { + "status": "PASS", + "variants": { + "thumbnail": { + "content_type": "image/jpeg", + "height": 4, + "name": "thumbnail", + "sha256": "a89b0af8a3c520ca079ed01f7a6579b39fdf0a593d5ff94b7a5128c194f3d065", + "size_bytes": 631, + "width": 8 + }, + "crop": { + "content_type": "image/jpeg", + "height": 128, + "name": "crop", + "sha256": "257edb1205654cf43e411c9fc333ee5544c2dc8753470ab667cdafbfeae4ff7c", + "size_bytes": 883, + "width": 128 + }, + "optimized": { + "content_type": "image/webp", + "height": 4, + "name": "optimized", + "sha256": "f4e61263caab4e749c64fc56091797191901d5affdb15ee2d828d905c429e950", + "size_bytes": 44, + "width": 8 + } + } + }, + "large": { + "status": "PASS", + "variants": { + "thumbnail": { + "content_type": "image/jpeg", + "height": 128, + "name": "thumbnail", + "sha256": "126971107bc487792a7b813c01e037c58140ab0a039d4c7fd0cdbd26a9d9d781", + "size_bytes": 1142, + "width": 256 + }, + "crop": { + "content_type": "image/jpeg", + "height": 128, + "name": "crop", + "sha256": "4e67b8bf92db7d696303a433822a79479ca6d8ac5ca6bf2c87e5f4514ed3fa4a", + "size_bytes": 886, + "width": 128 + }, + "optimized": { + "content_type": "image/webp", + "height": 640, + "name": "optimized", + "sha256": "8e5a8340d7a7b9c37e935e090bb34494e588b7534a6d091aba33aca5ba83a267", + "size_bytes": 1540, + "width": 1280 + } + } + }, + "boundary": { + "status": "PASS", + "variants": { + "thumbnail": { + "content_type": "image/jpeg", + "height": 160, + "name": "thumbnail", + "sha256": "d8d8358f529fd170984ee7926c54e9f6ead7d6a397fd6b803a6c20af197a2ce2", + "size_bytes": 1270, + "width": 256 + }, + "crop": { + "content_type": "image/jpeg", + "height": 128, + "name": "crop", + "sha256": "4e67b8bf92db7d696303a433822a79479ca6d8ac5ca6bf2c87e5f4514ed3fa4a", + "size_bytes": 886, + "width": 128 + }, + "optimized": { + "content_type": "image/webp", + "height": 800, + "name": "optimized", + "sha256": "a9bc2a93f44b92bf58d3c1a2f53390b21f0330b073411886eeccac3b12eb6769", + "size_bytes": 1904, + "width": 1280 + } + } + }, + "byte_limits_corrupt_pixel_limits": "PASS", + "private_bucket_limited_credentials": "PASS", + "worker_memory_bytes": { + "memory.current": 23502848, + "memory.max": 1610612736, + "memory.peak": 704679936 + } +} diff --git a/go.mod b/go.mod new file mode 100644 index 0000000..6e73275 --- /dev/null +++ b/go.mod @@ -0,0 +1,74 @@ +module mtusi-messenger + +go 1.26.0 + +require ( + github.com/google/uuid v1.6.0 + github.com/jackc/pgx/v5 v5.7.6 + github.com/minio/madmin-go/v3 v3.0.95 + github.com/minio/minio-go/v7 v7.0.95 + github.com/redis/go-redis/v9 v9.14.0 + go.opentelemetry.io/otel v1.38.0 + go.opentelemetry.io/otel/exporters/otlp/otlplog/otlploghttp v0.14.0 + go.opentelemetry.io/otel/exporters/otlp/otlpmetric/otlpmetrichttp v1.38.0 + go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp v1.38.0 + go.opentelemetry.io/otel/log v0.14.0 + go.opentelemetry.io/otel/metric v1.38.0 + go.opentelemetry.io/otel/sdk v1.38.0 + go.opentelemetry.io/otel/sdk/log v0.14.0 + go.opentelemetry.io/otel/sdk/metric v1.38.0 + go.opentelemetry.io/otel/trace v1.38.0 + golang.org/x/crypto v0.42.0 +) + +require ( + github.com/cenkalti/backoff/v5 v5.0.3 // indirect + github.com/cespare/xxhash/v2 v2.3.0 // indirect + github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f // indirect + github.com/dustin/go-humanize v1.0.1 // indirect + github.com/go-ini/ini v1.67.0 // indirect + github.com/go-logr/logr v1.4.3 // indirect + github.com/go-logr/stdr v1.2.2 // indirect + github.com/go-ole/go-ole v1.3.0 // indirect + github.com/goccy/go-json v0.10.5 // indirect + github.com/golang-jwt/jwt/v4 v4.5.0 // indirect + github.com/golang/protobuf v1.5.4 // indirect + github.com/grpc-ecosystem/grpc-gateway/v2 v2.27.2 // indirect + github.com/jackc/pgpassfile v1.0.0 // indirect + github.com/jackc/pgservicefile v0.0.0-20240606120523-5a60cdf6a761 // indirect + github.com/jackc/puddle/v2 v2.2.2 // indirect + github.com/klauspost/compress v1.18.0 // indirect + github.com/klauspost/cpuid/v2 v2.2.11 // indirect + github.com/lufia/plan9stats v0.0.0-20240909124753-873cd0166683 // indirect + github.com/matttproud/golang_protobuf_extensions v1.0.4 // indirect + github.com/minio/crc64nvme v1.0.2 // indirect + github.com/minio/md5-simd v1.1.2 // indirect + github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 // indirect + github.com/philhofer/fwd v1.2.0 // indirect + github.com/power-devops/perfstat v0.0.0-20240221224432-82ca36839d55 // indirect + github.com/prometheus/client_model v0.6.1 // indirect + github.com/prometheus/common v0.59.1 // indirect + github.com/prometheus/procfs v0.15.1 // indirect + github.com/prometheus/prom2json v1.4.0 // indirect + github.com/prometheus/prometheus v0.54.1 // indirect + github.com/rs/xid v1.6.0 // indirect + github.com/safchain/ethtool v0.4.1 // indirect + github.com/secure-io/sio-go v0.3.1 // indirect + github.com/shirou/gopsutil/v3 v3.24.5 // indirect + github.com/shoenig/go-m1cpu v0.1.6 // indirect + github.com/tinylib/msgp v1.3.0 // indirect + github.com/tklauser/go-sysconf v0.3.14 // indirect + github.com/tklauser/numcpus v0.8.0 // indirect + github.com/yusufpapurcu/wmi v1.2.4 // indirect + go.opentelemetry.io/auto/sdk v1.1.0 // indirect + go.opentelemetry.io/otel/exporters/otlp/otlptrace v1.38.0 // indirect + go.opentelemetry.io/proto/otlp v1.7.1 // indirect + golang.org/x/net v0.43.0 // indirect + golang.org/x/sync v0.17.0 // indirect + golang.org/x/sys v0.36.0 // indirect + golang.org/x/text v0.29.0 // indirect + google.golang.org/genproto/googleapis/api v0.0.0-20250825161204-c5933d9347a5 // indirect + google.golang.org/genproto/googleapis/rpc v0.0.0-20250825161204-c5933d9347a5 // indirect + google.golang.org/grpc v1.75.0 // indirect + google.golang.org/protobuf v1.36.8 // indirect +) diff --git a/go.sum b/go.sum new file mode 100644 index 0000000..563b14e --- /dev/null +++ b/go.sum @@ -0,0 +1,176 @@ +github.com/bsm/ginkgo/v2 v2.12.0 h1:Ny8MWAHyOepLGlLKYmXG4IEkioBysk6GpaRTLC8zwWs= +github.com/bsm/ginkgo/v2 v2.12.0/go.mod h1:SwYbGRRDovPVboqFv0tPTcG1sN61LM1Z4ARdbAV9g4c= +github.com/bsm/gomega v1.27.10 h1:yeMWxP2pV2fG3FgAODIY8EiRE3dy0aeFYt4l7wh6yKA= +github.com/bsm/gomega v1.27.10/go.mod h1:JyEr/xRbxbtgWNi8tIEVPUYZ5Dzef52k01W3YH0H+O0= +github.com/cenkalti/backoff/v5 v5.0.3 h1:ZN+IMa753KfX5hd8vVaMixjnqRZ3y8CuJKRKj1xcsSM= +github.com/cenkalti/backoff/v5 v5.0.3/go.mod h1:rkhZdG3JZukswDf7f0cwqPNk4K0sa+F97BxZthm/crw= +github.com/cespare/xxhash/v2 v2.3.0 h1:UL815xU9SqsFlibzuggzjXhog7bL6oX9BbNZnL2UFvs= +github.com/cespare/xxhash/v2 v2.3.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs= +github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38= +github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc h1:U9qPSI2PIWSS1VwoXQT9A3Wy9MM3WgvqSxFWenqJduM= +github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38= +github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f h1:lO4WD4F/rVNCu3HqELle0jiPLLBs70cWOduZpkS1E78= +github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f/go.mod h1:cuUVRXasLTGF7a8hSLbxyZXjz+1KgoB3wDUb6vlszIc= +github.com/dustin/go-humanize v1.0.1 h1:GzkhY7T5VNhEkwH0PVJgjz+fX1rhBrR7pRT3mDkpeCY= +github.com/dustin/go-humanize v1.0.1/go.mod h1:Mu1zIs6XwVuF/gI1OepvI0qD18qycQx+mFykh5fBlto= +github.com/go-ini/ini v1.67.0 h1:z6ZrTEZqSWOTyH2FlglNbNgARyHG8oLW9gMELqKr06A= +github.com/go-ini/ini v1.67.0/go.mod h1:ByCAeIL28uOIIG0E3PJtZPDL8WnHpFKFOtgjp+3Ies8= +github.com/go-logr/logr v1.2.2/go.mod h1:jdQByPbusPIv2/zmleS9BjJVeZ6kBagPoEUsqbVz/1A= +github.com/go-logr/logr v1.4.3 h1:CjnDlHq8ikf6E492q6eKboGOC0T8CDaOvkHCIg8idEI= +github.com/go-logr/logr v1.4.3/go.mod h1:9T104GzyrTigFIr8wt5mBrctHMim0Nb2HLGrmQ40KvY= +github.com/go-logr/stdr v1.2.2 h1:hSWxHoqTgW2S2qGc0LTAI563KZ5YKYRhT3MFKZMbjag= +github.com/go-logr/stdr v1.2.2/go.mod h1:mMo/vtBO5dYbehREoey6XUKy/eSumjCCveDpRre4VKE= +github.com/go-ole/go-ole v1.2.6/go.mod h1:pprOEPIfldk/42T2oK7lQ4v4JSDwmV0As9GaiUsvbm0= +github.com/go-ole/go-ole v1.3.0 h1:Dt6ye7+vXGIKZ7Xtk4s6/xVdGDQynvom7xCFEdWr6uE= +github.com/go-ole/go-ole v1.3.0/go.mod h1:5LS6F96DhAwUc7C+1HLexzMXY1xGRSryjyPPKW6zv78= +github.com/goccy/go-json v0.10.5 h1:Fq85nIqj+gXn/S5ahsiTlK3TmC85qgirsdTP/+DeaC4= +github.com/goccy/go-json v0.10.5/go.mod h1:oq7eo15ShAhp70Anwd5lgX2pLfOS3QCiwU/PULtXL6M= +github.com/golang-jwt/jwt/v4 v4.5.0 h1:7cYmW1XlMY7h7ii7UhUyChSgS5wUJEnm9uZVTGqOWzg= +github.com/golang-jwt/jwt/v4 v4.5.0/go.mod h1:m21LjoU+eqJr34lmDMbreY2eSTRJ1cv77w39/MY0Ch0= +github.com/golang/protobuf v1.2.0/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U= +github.com/golang/protobuf v1.5.4 h1:i7eJL8qZTpSEXOPTxNKhASYpMn+8e5Q6AdndVa1dWek= +github.com/golang/protobuf v1.5.4/go.mod h1:lnTiLA8Wa4RWRcIUkrtSVa5nRhsEGBg48fD6rSs7xps= +github.com/google/go-cmp v0.6.0/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY= +github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8= +github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU= +github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0= +github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo= +github.com/grpc-ecosystem/grpc-gateway/v2 v2.27.2 h1:8Tjv8EJ+pM1xP8mK6egEbD1OgnVTyacbefKhmbLhIhU= +github.com/grpc-ecosystem/grpc-gateway/v2 v2.27.2/go.mod h1:pkJQ2tZHJ0aFOVEEot6oZmaVEZcRme73eIFmhiVuRWs= +github.com/jackc/pgpassfile v1.0.0 h1:/6Hmqy13Ss2zCq62VdNG8tM1wchn8zjSGOBJ6icpsIM= +github.com/jackc/pgpassfile v1.0.0/go.mod h1:CEx0iS5ambNFdcRtxPj5JhEz+xB6uRky5eyVu/W2HEg= +github.com/jackc/pgservicefile v0.0.0-20240606120523-5a60cdf6a761 h1:iCEnooe7UlwOQYpKFhBabPMi4aNAfoODPEFNiAnClxo= +github.com/jackc/pgservicefile v0.0.0-20240606120523-5a60cdf6a761/go.mod h1:5TJZWKEWniPve33vlWYSoGYefn3gLQRzjfDlhSJ9ZKM= +github.com/jackc/pgx/v5 v5.7.6 h1:rWQc5FwZSPX58r1OQmkuaNicxdmExaEz5A2DO2hUuTk= +github.com/jackc/pgx/v5 v5.7.6/go.mod h1:aruU7o91Tc2q2cFp5h4uP3f6ztExVpyVv88Xl/8Vl8M= +github.com/jackc/puddle/v2 v2.2.2 h1:PR8nw+E/1w0GLuRFSmiioY6UooMp6KJv0/61nB7icHo= +github.com/jackc/puddle/v2 v2.2.2/go.mod h1:vriiEXHvEE654aYKXXjOvZM39qJ0q+azkZFrfEOc3H4= +github.com/klauspost/compress v1.18.0 h1:c/Cqfb0r+Yi+JtIEq73FWXVkRonBlf0CRNYc8Zttxdo= +github.com/klauspost/compress v1.18.0/go.mod h1:2Pp+KzxcywXVXMr50+X0Q/Lsb43OQHYWRCY2AiWywWQ= +github.com/klauspost/cpuid/v2 v2.0.1/go.mod h1:FInQzS24/EEf25PyTYn52gqo7WaD8xa0213Md/qVLRg= +github.com/klauspost/cpuid/v2 v2.2.11 h1:0OwqZRYI2rFrjS4kvkDnqJkKHdHaRnCm68/DY4OxRzU= +github.com/klauspost/cpuid/v2 v2.2.11/go.mod h1:hqwkgyIinND0mEev00jJYCxPNVRVXFQeu1XKlok6oO0= +github.com/lufia/plan9stats v0.0.0-20240909124753-873cd0166683 h1:7UMa6KCCMjZEMDtTVdcGu0B1GmmC7QJKiCCjyTAWQy0= +github.com/lufia/plan9stats v0.0.0-20240909124753-873cd0166683/go.mod h1:ilwx/Dta8jXAgpFYFvSWEMwxmbWXyiUHkd5FwyKhb5k= +github.com/matttproud/golang_protobuf_extensions v1.0.4 h1:mmDVorXM7PCGKw94cs5zkfA9PSy5pEvNWRP0ET0TIVo= +github.com/matttproud/golang_protobuf_extensions v1.0.4/go.mod h1:BSXmuO+STAnVfrANrmjBb36TMTDstsz7MSK+HVaYKv4= +github.com/minio/crc64nvme v1.0.2 h1:6uO1UxGAD+kwqWWp7mBFsi5gAse66C4NXO8cmcVculg= +github.com/minio/crc64nvme v1.0.2/go.mod h1:eVfm2fAzLlxMdUGc0EEBGSMmPwmXD5XiNRpnu9J3bvg= +github.com/minio/madmin-go/v3 v3.0.95 h1:KRy4NIpTLKrHcuTQbKrwKs3SvGoe+ZAHT2Pbl1zJxDc= +github.com/minio/madmin-go/v3 v3.0.95/go.mod h1:pMLdj9OtN0CANNs5tdm6opvOlDFfj0WhbztboZAjRWE= +github.com/minio/md5-simd v1.1.2 h1:Gdi1DZK69+ZVMoNHRXJyNcxrMA4dSxoYHZSQbirFg34= +github.com/minio/md5-simd v1.1.2/go.mod h1:MzdKDxYpY2BT9XQFocsiZf/NKVtR7nkE4RoEpN+20RM= +github.com/minio/minio-go/v7 v7.0.95 h1:ywOUPg+PebTMTzn9VDsoFJy32ZuARN9zhB+K3IYEvYU= +github.com/minio/minio-go/v7 v7.0.95/go.mod h1:wOOX3uxS334vImCNRVyIDdXX9OsXDm89ToynKgqUKlo= +github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 h1:C3w9PqII01/Oq1c1nUAm88MOHcQC9l5mIlSMApZMrHA= +github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822/go.mod h1:+n7T8mK8HuQTcFwEeznm/DIxMOiR9yIdICNftLE1DvQ= +github.com/philhofer/fwd v1.2.0 h1:e6DnBTl7vGY+Gz322/ASL4Gyp1FspeMvx1RNDoToZuM= +github.com/philhofer/fwd v1.2.0/go.mod h1:RqIHx9QI14HlwKwm98g9Re5prTQ6LdeRQn+gXJFxsJM= +github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4= +github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2 h1:Jamvg5psRIccs7FGNTlIRMkT8wgtp5eCXdBlqhYGL6U= +github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4= +github.com/power-devops/perfstat v0.0.0-20240221224432-82ca36839d55 h1:o4JXh1EVt9k/+g42oCprj/FisM4qX9L3sZB3upGN2ZU= +github.com/power-devops/perfstat v0.0.0-20240221224432-82ca36839d55/go.mod h1:OmDBASR4679mdNQnz2pUhc2G8CO2JrUAVFDRBDP/hJE= +github.com/prometheus/client_model v0.6.1 h1:ZKSh/rekM+n3CeS952MLRAdFwIKqeY8b62p8ais2e9E= +github.com/prometheus/client_model v0.6.1/go.mod h1:OrxVMOVHjw3lKMa8+x6HeMGkHMQyHDk9E3jmP2AmGiY= +github.com/prometheus/common v0.59.1 h1:LXb1quJHWm1P6wq/U824uxYi4Sg0oGvNeUm1z5dJoX0= +github.com/prometheus/common v0.59.1/go.mod h1:GpWM7dewqmVYcd7SmRaiWVe9SSqjf0UrwnYnpEZNuT0= +github.com/prometheus/procfs v0.15.1 h1:YagwOFzUgYfKKHX6Dr+sHT7km/hxC76UB0learggepc= +github.com/prometheus/procfs v0.15.1/go.mod h1:fB45yRUv8NstnjriLhBQLuOUt+WW4BsoGhij/e3PBqk= +github.com/prometheus/prom2json v1.4.0 h1:2AEOsd1ebqql/p9u0IWgCpUAteAAf9Lnf/SVyieqer4= +github.com/prometheus/prom2json v1.4.0/go.mod h1:DmcIMPspQD/fMyFCYti5qJJbuEnqDh3DGoooO0sgr4w= +github.com/prometheus/prometheus v0.54.1 h1:vKuwQNjnYN2/mDoWfHXDhAsz/68q/dQDb+YbcEqU7MQ= +github.com/prometheus/prometheus v0.54.1/go.mod h1:xlLByHhk2g3ycakQGrMaU8K7OySZx98BzeCR99991NY= +github.com/redis/go-redis/v9 v9.14.0 h1:u4tNCjXOyzfgeLN+vAZaW1xUooqWDqVEsZN0U01jfAE= +github.com/redis/go-redis/v9 v9.14.0/go.mod h1:huWgSWd8mW6+m0VPhJjSSQ+d6Nh1VICQ6Q5lHuCH/Iw= +github.com/rs/xid v1.6.0 h1:fV591PaemRlL6JfRxGDEPl69wICngIQ3shQtzfy2gxU= +github.com/rs/xid v1.6.0/go.mod h1:7XoLgs4eV+QndskICGsho+ADou8ySMSjJKDIan90Nz0= +github.com/safchain/ethtool v0.4.1 h1:S6mEleTADqgynileXoiapt/nKnatyR6bmIHoF+h2ADo= +github.com/safchain/ethtool v0.4.1/go.mod h1:XLLnZmy4OCRTkksP/UiMjij96YmIsBfmBQcs7H6tA48= +github.com/secure-io/sio-go v0.3.1 h1:dNvY9awjabXTYGsTF1PiCySl9Ltofk9GA3VdWlo7rRc= +github.com/secure-io/sio-go v0.3.1/go.mod h1:+xbkjDzPjwh4Axd07pRKSNriS9SCiYksWnZqdnfpQxs= +github.com/shirou/gopsutil/v3 v3.24.5 h1:i0t8kL+kQTvpAYToeuiVk3TgDeKOFioZO3Ztz/iZ9pI= +github.com/shirou/gopsutil/v3 v3.24.5/go.mod h1:bsoOS1aStSs9ErQ1WWfxllSeS1K5D+U30r2NfcubMVk= +github.com/shoenig/go-m1cpu v0.1.6 h1:nxdKQNcEB6vzgA2E2bvzKIYRuNj7XNJ4S/aRSwKzFtM= +github.com/shoenig/go-m1cpu v0.1.6/go.mod h1:1JJMcUBvfNwpq05QDQVAnx3gUHr9IYF7GNg9SUEw2VQ= +github.com/shoenig/test v0.6.4 h1:kVTaSd7WLz5WZ2IaoM0RSzRsUD+m8wRR+5qvntpn4LU= +github.com/shoenig/test v0.6.4/go.mod h1:byHiCGXqrVaflBLAMq/srcZIHynQPQgeyvkvXnjqq0k= +github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME= +github.com/stretchr/testify v1.3.0/go.mod h1:M5WIy9Dh21IEIfnGCwXGc5bZfKNJtfHm1UVUgZn+9EI= +github.com/stretchr/testify v1.7.0/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg= +github.com/stretchr/testify v1.11.1 h1:7s2iGBzp5EwR7/aIZr8ao5+dra3wiQyKjjFuvgVKu7U= +github.com/stretchr/testify v1.11.1/go.mod h1:wZwfW3scLgRK+23gO65QZefKpKQRnfz6sD981Nm4B6U= +github.com/tinylib/msgp v1.3.0 h1:ULuf7GPooDaIlbyvgAxBV/FI7ynli6LZ1/nVUNu+0ww= +github.com/tinylib/msgp v1.3.0/go.mod h1:ykjzy2wzgrlvpDCRc4LA8UXy6D8bzMSuAF3WD57Gok0= +github.com/tklauser/go-sysconf v0.3.14 h1:g5vzr9iPFFz24v2KZXs/pvpvh8/V9Fw6vQK5ZZb78yU= +github.com/tklauser/go-sysconf v0.3.14/go.mod h1:1ym4lWMLUOhuBOPGtRcJm7tEGX4SCYNEEEtghGG/8uY= +github.com/tklauser/numcpus v0.8.0 h1:Mx4Wwe/FjZLeQsK/6kt2EOepwwSl7SmJrK5bV/dXYgY= +github.com/tklauser/numcpus v0.8.0/go.mod h1:ZJZlAY+dmR4eut8epnzf0u/VwodKmryxR8txiloSqBE= +github.com/yusufpapurcu/wmi v1.2.4 h1:zFUKzehAFReQwLys1b/iSMl+JQGSCSjtVqQn9bBrPo0= +github.com/yusufpapurcu/wmi v1.2.4/go.mod h1:SBZ9tNy3G9/m5Oi98Zks0QjeHVDvuK0qfxQmPyzfmi0= +go.opentelemetry.io/auto/sdk v1.1.0 h1:cH53jehLUN6UFLY71z+NDOiNJqDdPRaXzTel0sJySYA= +go.opentelemetry.io/auto/sdk v1.1.0/go.mod h1:3wSPjt5PWp2RhlCcmmOial7AvC4DQqZb7a7wCow3W8A= +go.opentelemetry.io/otel v1.38.0 h1:RkfdswUDRimDg0m2Az18RKOsnI8UDzppJAtj01/Ymk8= +go.opentelemetry.io/otel v1.38.0/go.mod h1:zcmtmQ1+YmQM9wrNsTGV/q/uyusom3P8RxwExxkZhjM= +go.opentelemetry.io/otel/exporters/otlp/otlplog/otlploghttp v0.14.0 h1:QQqYw3lkrzwVsoEX0w//EhH/TCnpRdEenKBOOEIMjWc= +go.opentelemetry.io/otel/exporters/otlp/otlplog/otlploghttp v0.14.0/go.mod h1:gSVQcr17jk2ig4jqJ2DX30IdWH251JcNAecvrqTxH1s= +go.opentelemetry.io/otel/exporters/otlp/otlpmetric/otlpmetrichttp v1.38.0 h1:Oe2z/BCg5q7k4iXC3cqJxKYg0ieRiOqF0cecFYdPTwk= +go.opentelemetry.io/otel/exporters/otlp/otlpmetric/otlpmetrichttp v1.38.0/go.mod h1:ZQM5lAJpOsKnYagGg/zV2krVqTtaVdYdDkhMoX6Oalg= +go.opentelemetry.io/otel/exporters/otlp/otlptrace v1.38.0 h1:GqRJVj7UmLjCVyVJ3ZFLdPRmhDUp2zFmQe3RHIOsw24= +go.opentelemetry.io/otel/exporters/otlp/otlptrace v1.38.0/go.mod h1:ri3aaHSmCTVYu2AWv44YMauwAQc0aqI9gHKIcSbI1pU= +go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp v1.38.0 h1:aTL7F04bJHUlztTsNGJ2l+6he8c+y/b//eR0jjjemT4= +go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp v1.38.0/go.mod h1:kldtb7jDTeol0l3ewcmd8SDvx3EmIE7lyvqbasU3QC4= +go.opentelemetry.io/otel/log v0.14.0 h1:2rzJ+pOAZ8qmZ3DDHg73NEKzSZkhkGIua9gXtxNGgrM= +go.opentelemetry.io/otel/log v0.14.0/go.mod h1:5jRG92fEAgx0SU/vFPxmJvhIuDU9E1SUnEQrMlJpOno= +go.opentelemetry.io/otel/metric v1.38.0 h1:Kl6lzIYGAh5M159u9NgiRkmoMKjvbsKtYRwgfrA6WpA= +go.opentelemetry.io/otel/metric v1.38.0/go.mod h1:kB5n/QoRM8YwmUahxvI3bO34eVtQf2i4utNVLr9gEmI= +go.opentelemetry.io/otel/sdk v1.38.0 h1:l48sr5YbNf2hpCUj/FoGhW9yDkl+Ma+LrVl8qaM5b+E= +go.opentelemetry.io/otel/sdk v1.38.0/go.mod h1:ghmNdGlVemJI3+ZB5iDEuk4bWA3GkTpW+DOoZMYBVVg= +go.opentelemetry.io/otel/sdk/log v0.14.0 h1:JU/U3O7N6fsAXj0+CXz21Czg532dW2V4gG1HE/e8Zrg= +go.opentelemetry.io/otel/sdk/log v0.14.0/go.mod h1:imQvII+0ZylXfKU7/wtOND8Hn4OpT3YUoIgqJVksUkM= +go.opentelemetry.io/otel/sdk/log/logtest v0.14.0 h1:Ijbtz+JKXl8T2MngiwqBlPaHqc4YCaP/i13Qrow6gAM= +go.opentelemetry.io/otel/sdk/log/logtest v0.14.0/go.mod h1:dCU8aEL6q+L9cYTqcVOk8rM9Tp8WdnHOPLiBgp0SGOA= +go.opentelemetry.io/otel/sdk/metric v1.38.0 h1:aSH66iL0aZqo//xXzQLYozmWrXxyFkBJ6qT5wthqPoM= +go.opentelemetry.io/otel/sdk/metric v1.38.0/go.mod h1:dg9PBnW9XdQ1Hd6ZnRz689CbtrUp0wMMs9iPcgT9EZA= +go.opentelemetry.io/otel/trace v1.38.0 h1:Fxk5bKrDZJUH+AMyyIXGcFAPah0oRcT+LuNtJrmcNLE= +go.opentelemetry.io/otel/trace v1.38.0/go.mod h1:j1P9ivuFsTceSWe1oY+EeW3sc+Pp42sO++GHkg4wwhs= +go.opentelemetry.io/proto/otlp v1.7.1 h1:gTOMpGDb0WTBOP8JaO72iL3auEZhVmAQg4ipjOVAtj4= +go.opentelemetry.io/proto/otlp v1.7.1/go.mod h1:b2rVh6rfI/s2pHWNlB7ILJcRALpcNDzKhACevjI+ZnE= +go.uber.org/goleak v1.3.0 h1:2K3zAYmnTNqV73imy9J1T3WC+gmCePx2hEGkimedGto= +go.uber.org/goleak v1.3.0/go.mod h1:CoHD4mav9JJNrW/WLlf7HGZPjdw8EucARQHekz1X6bE= +golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w= +golang.org/x/crypto v0.0.0-20200302210943-78000ba7a073/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto= +golang.org/x/crypto v0.42.0 h1:chiH31gIWm57EkTXpwnqf8qeuMUi0yekh6mT2AvFlqI= +golang.org/x/crypto v0.42.0/go.mod h1:4+rDnOTJhQCx2q7/j6rAN5XDw8kPjeaXEUR2eL94ix8= +golang.org/x/net v0.0.0-20190404232315-eb5bcb51f2a3/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg= +golang.org/x/net v0.43.0 h1:lat02VYK2j4aLzMzecihNvTlJNQUq316m2Mr9rnM6YE= +golang.org/x/net v0.43.0/go.mod h1:vhO1fvI4dGsIjh73sWfUVjj3N7CA9WkKJNQm2svM6Jg= +golang.org/x/sync v0.0.0-20181221193216-37e7f081c4d4/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM= +golang.org/x/sync v0.17.0 h1:l60nONMj9l5drqw6jlhIELNv9I0A4OFgRsG9k2oT9Ug= +golang.org/x/sync v0.17.0/go.mod h1:9KTHXmSnoGruLpwFjVSX0lNNA75CykiMECbovNTZqGI= +golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY= +golang.org/x/sys v0.0.0-20190412213103-97732733099d/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs= +golang.org/x/sys v0.0.0-20190916202348-b4ddaad3f8a3/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs= +golang.org/x/sys v0.0.0-20200302150141-5c8b2ff67527/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs= +golang.org/x/sys v0.0.0-20201204225414-ed752295db88/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs= +golang.org/x/sys v0.1.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg= +golang.org/x/sys v0.21.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA= +golang.org/x/sys v0.36.0 h1:KVRy2GtZBrk1cBYA7MKu5bEZFxQk4NIDV6RLVcC8o0k= +golang.org/x/sys v0.36.0/go.mod h1:OgkHotnGiDImocRcuBABYBEXf8A9a87e/uXjp9XT3ks= +golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ= +golang.org/x/text v0.29.0 h1:1neNs90w9YzJ9BocxfsQNHKuAT4pkghyXc4nhZ6sJvk= +golang.org/x/text v0.29.0/go.mod h1:7MhJOA9CD2qZyOKYazxdYMF85OwPdEr9jTtBpO7ydH4= +gonum.org/v1/gonum v0.16.0 h1:5+ul4Swaf3ESvrOnidPp4GZbzf0mxVQpDCYUQE7OJfk= +gonum.org/v1/gonum v0.16.0/go.mod h1:fef3am4MQ93R2HHpKnLk4/Tbh/s0+wqD5nfa6Pnwy4E= +google.golang.org/genproto/googleapis/api v0.0.0-20250825161204-c5933d9347a5 h1:BIRfGDEjiHRrk0QKZe3Xv2ieMhtgRGeLcZQ0mIVn4EY= +google.golang.org/genproto/googleapis/api v0.0.0-20250825161204-c5933d9347a5/go.mod h1:j3QtIyytwqGr1JUDtYXwtMXWPKsEa5LtzIFN1Wn5WvE= +google.golang.org/genproto/googleapis/rpc v0.0.0-20250825161204-c5933d9347a5 h1:eaY8u2EuxbRv7c3NiGK0/NedzVsCcV6hDuU5qPX5EGE= +google.golang.org/genproto/googleapis/rpc v0.0.0-20250825161204-c5933d9347a5/go.mod h1:M4/wBTSeyLxupu3W3tJtOgB14jILAS/XWPSSa3TAlJc= +google.golang.org/grpc v1.75.0 h1:+TW+dqTd2Biwe6KKfhE5JpiYIBWq865PhKGSXiivqt4= +google.golang.org/grpc v1.75.0/go.mod h1:JtPAzKiq4v1xcAB2hydNlWI2RnF85XXcV0mhKXr2ecQ= +google.golang.org/protobuf v1.36.8 h1:xHScyCOEuuwZEc6UtSOvPbAT4zRh0xcNRYekJwfqyMc= +google.golang.org/protobuf v1.36.8/go.mod h1:fuxRtAxBytpl4zzqUh6/eyUujkJdNiuEkXntxiD/uRU= +gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0= +gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM= +gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA= +gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM= diff --git a/infra/Dockerfile.minio b/infra/Dockerfile.minio new file mode 100644 index 0000000..4be57c1 --- /dev/null +++ b/infra/Dockerfile.minio @@ -0,0 +1,16 @@ +# Build MinIO from a pinned commit of the official upstream source. +FROM golang:1.26.8-alpine3.23@sha256:a8fa79c5bd40d880b52bd3b6d7669ecdcfd00e85facdd427d279efb5ddd79cb1 AS build +ADD --checksum=sha256:be6d0bd3696c3a13a35f02d3a0280b64319c67918b4501c5c3d87f96d000085c https://codeload.github.com/minio/minio/tar.gz/refs/tags/RELEASE.2025-10-15T17-29-55Z /source.tar.gz +WORKDIR /src +RUN tar -xzf /source.tar.gz --strip-components=1 -C /src +RUN --mount=type=cache,target=/go/pkg/mod --mount=type=cache,target=/root/.cache/go-build \ + CGO_ENABLED=0 GOMAXPROCS=2 go build -p 2 -trimpath -buildvcs=false \ + -ldflags='-s -w -X github.com/minio/minio/cmd.Version=2025-10-15T17:29:55Z -X github.com/minio/minio/cmd.ReleaseTag=RELEASE.2025-10-15T17-29-55Z -X github.com/minio/minio/cmd.CommitID=9e49d5e7a648f00e26f2246f4dc28e6b07f8c84a -X github.com/minio/minio/cmd.CopyrightYear=2025' \ + -o /minio . + +FROM busybox:1.37.0-musl +COPY --from=build /minio /minio +COPY --from=build /src/LICENSE /licenses/MinIO-LICENSE +COPY --from=build /etc/ssl/certs/ca-certificates.crt /etc/ssl/certs/ca-certificates.crt +RUN mkdir -p /data +ENTRYPOINT ["/minio"] diff --git a/infra/Dockerfile.worker b/infra/Dockerfile.worker new file mode 100644 index 0000000..1b4674e --- /dev/null +++ b/infra/Dockerfile.worker @@ -0,0 +1,8 @@ +FROM application AS app_binary +FROM python:3.13.7-slim-bookworm +RUN pip install --no-cache-dir Pillow==12.0.0 +COPY --from=app_binary /messenger /messenger +COPY worker/process_image.py /opt/process_image.py +USER 65532:65532 +ENTRYPOINT ["/messenger"] +CMD ["worker"] diff --git a/infra/health-aware.sh b/infra/health-aware.sh new file mode 100644 index 0000000..4ce2da8 --- /dev/null +++ b/infra/health-aware.sh @@ -0,0 +1,43 @@ +#!/bin/sh +set -eu + +upstream=/run/nginx-upstream.conf + +write_servers() { + if [ "$1" = 1 ]; then + printf '%s\n' 'server api1:8080 resolve max_fails=0;' + fi + if [ "$2" = 1 ]; then + printf '%s\n' 'server api2:8080 resolve max_fails=0;' + fi +} + +# With no healthy replica, keep both reachable so API can return its normal +# 401/503 JSON responses instead of Nginx reporting "no live upstreams". +write_servers 1 1 > "$upstream" + +monitor() { + sleep 2 + while :; do + ready1=0 + ready2=0 + wget -q -T 2 -O /dev/null http://api1:8080/health/ready && ready1=1 || true + wget -q -T 2 -O /dev/null http://api2:8080/health/ready && ready2=1 || true + if [ "$ready1" = 0 ] && [ "$ready2" = 0 ]; then + ready1=1 + ready2=1 + fi + next=$(mktemp /run/nginx-upstream.XXXXXX) + write_servers "$ready1" "$ready2" > "$next" + if cmp -s "$next" "$upstream"; then + rm -f "$next" + else + mv "$next" "$upstream" + nginx -t >/dev/null 2>&1 && nginx -s reload + fi + sleep 1 + done +} + +monitor & +exec nginx -g 'daemon off;' diff --git a/infra/nginx.conf b/infra/nginx.conf new file mode 100644 index 0000000..6cd7413 --- /dev/null +++ b/infra/nginx.conf @@ -0,0 +1,35 @@ +events {} + +http { + access_log /dev/stdout; + error_log /dev/stderr notice; + + resolver 127.0.0.11 valid=2s ipv6=off; + + upstream messenger_api { + zone messenger_api 64k; + include /run/nginx-upstream.conf; + keepalive 16; + } + + server { + listen 8080; + client_max_body_size 0; + client_body_timeout 20s; + + location / { + proxy_request_buffering off; + proxy_buffering off; + proxy_pass http://messenger_api; + proxy_http_version 1.1; + proxy_set_header Connection ""; + proxy_set_header Host $host; + proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; + proxy_connect_timeout 500ms; + proxy_send_timeout 65s; + proxy_read_timeout 65s; + proxy_next_upstream error timeout http_502 http_503 http_504; + proxy_next_upstream_tries 2; + } + } +} diff --git a/infra/otel/README.md b/infra/otel/README.md index c46fe89..14afd0d 100644 --- a/infra/otel/README.md +++ b/infra/otel/README.md @@ -1,7 +1,12 @@ -# Конфигурация наблюдаемости — реализовать +# Наблюдаемость лабораторной 6 -Добавьте `collector.yaml` с OTLP receivers, processors и exporters выбранной версии Collector. Укажите явные bind endpoints, внутренние сети, memory limits/batch и поведение при backpressure по уровню. Сопоставьте все три pipeline с реально развёрнутыми backend. +- `collector.yaml` — OTLP receiver и три pipeline, memory limiter, batch, bounded queues/retry. +- `prometheus.yaml`, `alerts.yaml` — scrape, SLI и два локальных alert rules. +- `loki.yaml` — OTLP logs с structured metadata и persistent filesystem. +- `tempo.yaml` — OTLP traces, WAL и persistent blocks. +- `grafana/provisioning` — три datasource и автоматический импорт dashboard. +- `grafana/dashboards/messenger.json` — импортируемый dashboard с реальными PromQL-запросами. +- [Сигналы и единицы](signals.md). +- [Runbook и SLO](runbook.md). -Добавьте конфигурации storage, Grafana provisioning/datasources, provisioning/dashboards и импортируемые dashboard JSON. На 4 — alert rules и runbooks. Пустой dashboard или debug exporter, который только печатает данные в stdout, не заменяют хранилища/доказательства из `contracts/observability.md`. - -Не публикуйте OTLP/metrics/административные ports в общую сеть без необходимости. В lab7 реализуйте TLS/mTLS по trust matrix; экспорт Authorization и message bodies запрещён. +Запуск и проверки — в корневом REPORT.md. Публикуется только Grafana на localhost; OTLP/Prometheus/Loki/Tempo ports остаются внутри Compose. diff --git a/infra/otel/alerts.yaml b/infra/otel/alerts.yaml new file mode 100644 index 0000000..b12f580 --- /dev/null +++ b/infra/otel/alerts.yaml @@ -0,0 +1,23 @@ +groups: + - name: messenger-slo + rules: + - record: messenger:api_error_ratio:rate1m + expr: (sum(rate(messenger_http_requests_total{service_name="messenger-api",route!~"/health/.*",status_class="5xx"}[1m])) or vector(0)) / clamp_min(sum(rate(messenger_http_requests_total{service_name="messenger-api",route!~"/health/.*"}[1m])), 0.001) + - record: messenger:api_availability:rate5m + expr: 1 - (sum(rate(messenger_http_requests_total{service_name="messenger-api",route!~"/health/.*",status_class="5xx"}[5m])) or vector(0)) / clamp_min(sum(rate(messenger_http_requests_total{service_name="messenger-api",route!~"/health/.*"}[5m])), 0.001) + - alert: MessengerHighErrorRate + expr: messenger:api_error_ratio:rate1m > 0.05 + for: 10s + labels: + severity: warning + annotations: + summary: Business API 5xx ratio exceeds 5% over 1 minute + description: Inspect database spans, DB availability and pool panels; see infra/otel/runbook.md. + - alert: MessengerQueueLag + expr: max(messenger_queue_oldest_age_seconds) > 10 and max(messenger_queue_depth) > 0 + for: 10s + labels: + severity: warning + annotations: + summary: Oldest queued job exceeds 10 seconds + description: Compare worker busy/capacity, job outcomes and queue age; see infra/otel/runbook.md. diff --git a/infra/otel/collector.yaml b/infra/otel/collector.yaml new file mode 100644 index 0000000..f068bd7 --- /dev/null +++ b/infra/otel/collector.yaml @@ -0,0 +1,74 @@ +receivers: + otlp: + protocols: + http: + endpoint: 0.0.0.0:4318 +processors: + memory_limiter: + check_interval: 1s + limit_mib: 128 + spike_limit_mib: 32 + batch: + timeout: 1s + send_batch_size: 64 + send_batch_max_size: 128 +exporters: + prometheus: + endpoint: 0.0.0.0:8889 + resource_to_telemetry_conversion: + enabled: true + otlphttp/tempo: + endpoint: http://tempo:4318 + timeout: 2s + sending_queue: + enabled: true + num_consumers: 2 + queue_size: 32 + block_on_overflow: false + retry_on_failure: + enabled: true + initial_interval: 1s + max_interval: 2s + max_elapsed_time: 5s + otlphttp/loki: + endpoint: http://loki:3100/otlp + timeout: 2s + sending_queue: + enabled: true + num_consumers: 2 + queue_size: 32 + block_on_overflow: false + retry_on_failure: + enabled: true + initial_interval: 1s + max_interval: 2s + max_elapsed_time: 5s +extensions: + health_check: + endpoint: 0.0.0.0:13133 +service: + extensions: [health_check] + telemetry: + logs: + encoding: json + metrics: + level: detailed + readers: + - pull: + exporter: + prometheus: + host: 0.0.0.0 + port: 8888 + pipelines: + metrics: + receivers: [otlp] + processors: [memory_limiter, batch] + exporters: [prometheus] + logs: + receivers: [otlp] + processors: [memory_limiter, batch] + exporters: [otlphttp/loki] + traces: + receivers: [otlp] + processors: [memory_limiter, batch] + exporters: [otlphttp/tempo] diff --git a/infra/otel/grafana/dashboards/messenger.json b/infra/otel/grafana/dashboards/messenger.json new file mode 100644 index 0000000..fe7327f --- /dev/null +++ b/infra/otel/grafana/dashboards/messenger.json @@ -0,0 +1,669 @@ +{ + "uid": "mtusi-messenger", + "title": "MTUSI Messenger — lab6", + "schemaVersion": 39, + "version": 1, + "refresh": "5s", + "tags": [ + "mtusi", + "lab6" + ], + "time": { + "from": "now-15m", + "to": "now" + }, + "panels": [ + { + "id": 1, + "title": "API RPS (business routes)", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 0, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "sum by (service_instance_id) (rate(messenger_http_requests_total{service_name=\"messenger-api\",route!~\"/health/.*\"}[1m]))", + "legendFormat": "{{service_instance_id}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "reqps" + }, + "overrides": [] + } + }, + { + "id": 2, + "title": "Business API error ratio: 5xx / all", + "type": "timeseries", + "gridPos": { + "x": 12, + "y": 0, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "messenger:api_error_ratio:rate1m", + "legendFormat": "5xx ratio" + } + ], + "fieldConfig": { + "defaults": { + "unit": "percentunit" + }, + "overrides": [] + } + }, + { + "id": 3, + "title": "API p50 from histogram", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 8, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "histogram_quantile(0.5, sum by (le) (rate(messenger_http_duration_seconds_bucket{service_name=\"messenger-api\",route!~\"/health/.*\"}[1m])))", + "legendFormat": "p50" + } + ], + "fieldConfig": { + "defaults": { + "unit": "s" + }, + "overrides": [] + } + }, + { + "id": 4, + "title": "API p95 from histogram", + "type": "timeseries", + "gridPos": { + "x": 12, + "y": 8, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "histogram_quantile(0.95, sum by (le) (rate(messenger_http_duration_seconds_bucket{service_name=\"messenger-api\",route!~\"/health/.*\"}[1m])))", + "legendFormat": "p95" + } + ], + "fieldConfig": { + "defaults": { + "unit": "s" + }, + "overrides": [] + } + }, + { + "id": 5, + "title": "API p99 from histogram", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 16, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "histogram_quantile(0.99, sum by (le) (rate(messenger_http_duration_seconds_bucket{service_name=\"messenger-api\",route!~\"/health/.*\"}[1m])))", + "legendFormat": "p99" + } + ], + "fieldConfig": { + "defaults": { + "unit": "s" + }, + "overrides": [] + } + }, + { + "id": 6, + "title": "Worker job outcomes", + "type": "timeseries", + "gridPos": { + "x": 12, + "y": 16, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "sum by (outcome) (rate(messenger_jobs_processed_total[1m]))", + "legendFormat": "{{outcome}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "ops" + }, + "overrides": [] + } + }, + { + "id": 7, + "title": "Worker p95 processing duration", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 24, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "histogram_quantile(0.95, sum by (le) (rate(messenger_job_duration_seconds_bucket[5m])))", + "legendFormat": "p95" + } + ], + "fieldConfig": { + "defaults": { + "unit": "s" + }, + "overrides": [] + } + }, + { + "id": 8, + "title": "End-to-end p95 time to ready", + "type": "timeseries", + "gridPos": { + "x": 12, + "y": 24, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "histogram_quantile(0.95, sum by (le) (rate(messenger_time_to_ready_seconds_bucket[5m])))", + "legendFormat": "p95" + } + ], + "fieldConfig": { + "defaults": { + "unit": "s" + }, + "overrides": [] + } + }, + { + "id": 9, + "title": "Queue depth", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 32, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "max(messenger_queue_depth)", + "legendFormat": "queued" + } + ], + "fieldConfig": { + "defaults": { + "unit": "short" + }, + "overrides": [] + } + }, + { + "id": 10, + "title": "Oldest queued job age", + "type": "timeseries", + "gridPos": { + "x": 12, + "y": 32, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "max(messenger_queue_oldest_age_seconds)", + "legendFormat": "oldest" + } + ], + "fieldConfig": { + "defaults": { + "unit": "s" + }, + "overrides": [] + } + }, + { + "id": 11, + "title": "DB pool used / waiting", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 40, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "messenger_db_pool_used", + "legendFormat": "used {{service_instance_id}}" + }, + { + "refId": "B", + "expr": "messenger_db_pool_waiting", + "legendFormat": "waiting {{service_instance_id}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "short" + }, + "overrides": [] + } + }, + { + "id": 12, + "title": "DB pool wait rate", + "type": "timeseries", + "gridPos": { + "x": 12, + "y": 40, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "sum by (service_instance_id) (rate(messenger_db_pool_acquire_wait_total[1m]))", + "legendFormat": "{{service_instance_id}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "ops" + }, + "overrides": [] + } + }, + { + "id": 13, + "title": "Worker saturation: busy / capacity", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 48, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "messenger_worker_busy / messenger_worker_capacity", + "legendFormat": "{{service_instance_id}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "percentunit" + }, + "overrides": [] + } + }, + { + "id": 14, + "title": "Worker capacity", + "type": "timeseries", + "gridPos": { + "x": 12, + "y": 48, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "messenger_worker_capacity", + "legendFormat": "{{service_instance_id}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "short" + }, + "overrides": [] + } + }, + { + "id": 15, + "title": "Container CPU time rate (includes decoder)", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 56, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "rate(messenger_container_cpu_seconds_total[1m])", + "legendFormat": "{{service_instance_id}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "cores" + }, + "overrides": [] + } + }, + { + "id": 16, + "title": "Container memory", + "type": "timeseries", + "gridPos": { + "x": 12, + "y": 56, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "messenger_container_memory_bytes", + "legendFormat": "{{service_instance_id}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "bytes" + }, + "overrides": [] + } + }, + { + "id": 17, + "title": "Database available", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 64, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "messenger_db_available", + "legendFormat": "{{service_instance_id}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "short" + }, + "overrides": [] + } + }, + { + "id": 18, + "title": "SLI availability, rolling 5m (SLO 99%)", + "type": "timeseries", + "gridPos": { + "x": 12, + "y": 64, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "messenger:api_availability:rate5m", + "legendFormat": "availability" + } + ], + "fieldConfig": { + "defaults": { + "unit": "percentunit" + }, + "overrides": [] + } + }, + { + "id": 19, + "title": "Local alert states", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 72, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "ALERTS{alertname=~\"Messenger.*\"}", + "legendFormat": "{{alertname}} {{alertstate}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "short" + }, + "overrides": [] + } + }, + { + "id": 20, + "title": "Collector queued export batches", + "type": "timeseries", + "gridPos": { + "x": 12, + "y": 72, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "otelcol_exporter_queue_size", + "legendFormat": "{{exporter}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "short" + }, + "overrides": [] + } + }, + { + "id": 21, + "title": "Collector failed trace exports", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 80, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "otelcol_exporter_send_failed_spans_total", + "legendFormat": "{{exporter}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "short" + }, + "overrides": [] + } + }, + { + "id": 22, + "title": "SDK exporter errors", + "type": "timeseries", + "gridPos": { + "x": 12, + "y": 80, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "messenger_telemetry_export_errors", + "legendFormat": "{{service_instance_id}}" + } + ], + "fieldConfig": { + "defaults": { + "unit": "short" + }, + "overrides": [] + } + }, + { + "id": 23, + "title": "Worker heartbeat age (stale if stopped)", + "type": "timeseries", + "gridPos": { + "x": 0, + "y": 88, + "w": 12, + "h": 8 + }, + "datasource": { + "type": "prometheus", + "uid": "prometheus" + }, + "targets": [ + { + "refId": "A", + "expr": "time() - messenger_process_last_observed_seconds{service_name=\"messenger-worker\"}", + "legendFormat": "worker heartbeat age" + } + ], + "fieldConfig": { + "defaults": { + "unit": "s" + }, + "overrides": [] + } + } + ], + "annotations": { + "list": [] + } +} diff --git a/infra/otel/grafana/provisioning/dashboards/provider.yaml b/infra/otel/grafana/provisioning/dashboards/provider.yaml new file mode 100644 index 0000000..89cc130 --- /dev/null +++ b/infra/otel/grafana/provisioning/dashboards/provider.yaml @@ -0,0 +1,11 @@ +apiVersion: 1 +providers: + - name: Messenger + orgId: 1 + folder: MTUSI + type: file + disableDeletion: false + editable: false + updateIntervalSeconds: 10 + options: + path: /var/lib/grafana/dashboards diff --git a/infra/otel/grafana/provisioning/datasources/sources.yaml b/infra/otel/grafana/provisioning/datasources/sources.yaml new file mode 100644 index 0000000..6e29b48 --- /dev/null +++ b/infra/otel/grafana/provisioning/datasources/sources.yaml @@ -0,0 +1,36 @@ +apiVersion: 1 +datasources: + - name: Prometheus + uid: prometheus + type: prometheus + access: proxy + url: http://prometheus:9090 + isDefault: true + editable: false + - name: Loki + uid: loki + type: loki + access: proxy + url: http://loki:3100 + editable: false + jsonData: + derivedFields: + - name: TraceID + matcherRegex: '"trace_id":"([a-f0-9]{32})"' + datasourceUid: tempo + url: '$${__value.raw}' + - name: Tempo + uid: tempo + type: tempo + access: proxy + url: http://tempo:3200 + editable: false + jsonData: + tracesToLogsV2: + datasourceUid: loki + spanStartTimeShift: '-1m' + spanEndTimeShift: '1m' + filterByTraceID: true + filterBySpanID: false + serviceMap: + datasourceUid: prometheus diff --git a/infra/otel/loki.yaml b/infra/otel/loki.yaml new file mode 100644 index 0000000..d2cbeb1 --- /dev/null +++ b/infra/otel/loki.yaml @@ -0,0 +1,30 @@ +auth_enabled: false +server: + http_listen_port: 3100 +common: + instance_addr: 127.0.0.1 + path_prefix: /loki + replication_factor: 1 + ring: + kvstore: + store: inmemory + storage: + filesystem: + chunks_directory: /loki/chunks + rules_directory: /loki/rules +schema_config: + configs: + - from: 2024-01-01 + store: tsdb + object_store: filesystem + schema: v13 + index: + prefix: index_ + period: 24h +limits_config: + allow_structured_metadata: true + retention_period: 24h +compactor: + working_directory: /loki/compactor + retention_enabled: true + delete_request_store: filesystem diff --git a/infra/otel/prometheus.yaml b/infra/otel/prometheus.yaml new file mode 100644 index 0000000..4fd5a97 --- /dev/null +++ b/infra/otel/prometheus.yaml @@ -0,0 +1,12 @@ +global: + scrape_interval: 3s + evaluation_interval: 5s +rule_files: + - /etc/prometheus/alerts.yaml +scrape_configs: + - job_name: messenger + static_configs: + - targets: [otel-collector:8889] + - job_name: collector + static_configs: + - targets: [otel-collector:8888] diff --git a/infra/otel/runbook.md b/infra/otel/runbook.md new file mode 100644 index 0000000..843c3ba --- /dev/null +++ b/infra/otel/runbook.md @@ -0,0 +1,42 @@ +# Локальная диагностика + +Открыть Grafana → MTUSI → **MTUSI Messenger — lab6**. В Explore доступны Prometheus, Loki и Tempo. Пароль admin хранится локально в `secrets/grafana_admin_password`, в Git его нет. + +## MessengerQueueLag + +Условие: oldest queued job > 10 s при depth > 0 в течение 10 s. Интервал оценки — 5 s, scrape — 3 s. + +1. Посмотреть Queue depth и Oldest queued job age: растут ли они вместе. +2. Посмотреть Worker heartbeat age и busy/capacity. Старый heartbeat при busy=0 означает отсутствие свежих наблюдений worker, а не доказательство его здоровья. +3. Найти `job.published` в Loki и trace upload в Tempo. Если producer есть, а нового consumer нет, искать недоступный worker/доставку. Если consumer есть с error, посмотреть его S3/DB child spans. +4. После устранения причины наблюдать jobs outcome=success, снижение depth/age, свежий heartbeat и resolved (alert исчезает из активных). + +Для демонстрации: `docker compose --env-file .env stop worker`, загрузить изображение; восстановить `docker compose --env-file .env start worker`. Проверки автоматизированы в `student-tests/verify_incidents.py`. + +## MessengerHighErrorRate + +Условие: 5xx / все бизнес-запросы > 5% за 1 минуту в течение 10 s. Health исключены. Ошибки клиента 400/401/404 не входят в числитель. + +1. Посмотреть Business API error ratio и API percentiles. +2. По `X-Trace-Id` ответа открыть trace в Tempo. В Loki фильтровать `trace_id` или `request_id`. +3. Для отказа PostgreSQL видны error spans `db.pool.acquire`/`db.select`, `messenger_db_available=0` и ожидание пула. Длительность HTTP без DB error сама по себе не доказывает отказ БД. +4. Восстановить PostgreSQL. Проверить успешные запросы, DB availability=1 и падение error ratio в движущемся окне. Resolved может наступить после обновления окна и нескольких scrape/evaluation. + +## Tempo недоступен + +1. API продолжает записи; проверить бизнес-операцию, а не только `/health/live`. +2. В Prometheus посмотреть `otelcol_exporter_send_failed_spans_total{exporter="otlphttp/tempo"}` и `otelcol_exporter_queue_size`. +3. Collector queue ограничена 32 batch requests, retry — 5 s; SDK queue — 256 spans/log records, exporter timeout — 2 s. Ожидание backend не переносится в HTTP handler. +4. При переполнении/исчерпании retry данные telemetry могут быть потеряны. В JSON-логах Collector финальная потеря содержит `dropped_items`. Счётчик failed sends не следует считать числом уникальных потерянных трасс. +5. После восстановления проверить свежий trace. Старые потерянные spans автоматически не восстанавливаются из бизнес-БД. + +## SLI/SLO + +SLI: доля бизнес HTTP-запросов без 5xx за скользящие **5 минут**. SLO: **≥99%**. + +```promql +1 - (sum(rate(messenger_http_requests_total{service_name="messenger-api",route!~"/health/.*",status_class="5xx"}[5m])) or vector(0)) + / clamp_min(sum(rate(messenger_http_requests_total{service_name="messenger-api",route!~"/health/.*"}[5m])), 0.001) +``` + +SLO оценивается на окне с реальным трафиком; при отсутствии запросов правило не должно порождать ошибку деления. Alert использует более короткое 1m окно и порог 5% для локальной диагностики. Это не burn-rate исследование оценки 5. diff --git a/infra/otel/signals.md b/infra/otel/signals.md new file mode 100644 index 0000000..e7fca89 --- /dev/null +++ b/infra/otel/signals.md @@ -0,0 +1,53 @@ +# Схема сигналов v1 + +SDK: OpenTelemetry Go traces/metrics **1.38.0**, logs **0.14.0**. Resource schema URL — OTel semantic conventions **1.37.0**; прикладные имена и labels ниже зафиксированы собственной схемой `mtusi-messenger/telemetry-v1`. + +Resources: `service.name` = messenger-api / messenger-worker / messenger-dispatcher; `service.version=lab6-v1`, `service.instance.id` = api-1/api-2/worker-1/dispatcher-1, `deployment.environment.name=lab`. Collector превращает resource attrs в Prometheus labels, например service_name. + +| Prometheus metric | Тип, единица | Labels | +|---|---|---| +| messenger_http_requests_total | counter, requests | method, route template, status_class | +| messenger_http_duration_seconds_bucket/count/sum | histogram, seconds | те же | +| messenger_jobs_processed_total | counter, attempts | job_type, outcome | +| messenger_job_duration_seconds_bucket/count/sum | histogram, seconds | job_type, outcome | +| messenger_time_to_ready_seconds_bucket/count/sum | histogram, seconds | job_type | +| messenger_queue_depth | gauge, jobs | queue; только dispatcher | +| messenger_queue_oldest_age_seconds | gauge, seconds | queue; только dispatcher | +| messenger_db_pool_used/idle/waiting/capacity | gauge, connections/acquires | resource instance | +| messenger_db_pool_acquire_wait_total | counter, empty pool acquires | resource instance | +| messenger_db_pool_acquire_wait_duration_seconds_total | counter, cumulative seconds | resource instance | +| messenger_worker_busy/capacity | gauge, attempts/slots | worker instance | +| messenger_process_cpu_seconds_total | counter, Go process CPU seconds | instance | +| messenger_container_cpu_seconds_total | counter, cgroup CPU seconds including decoder | instance | +| messenger_process_memory_bytes | gauge, Go runtime reserved memory (Sys), не RSS | instance | +| messenger_container_memory_bytes/limit_bytes | gauge, cgroup bytes | instance | +| messenger_process_last_observed_seconds | gauge, Unix timestamp | instance | +| messenger_db_available | gauge, 0/1 | dispatcher | +| messenger_telemetry_export_errors | gauge, cumulative SDK export errors | instance | + +`waiting` — незавершённые Acquire; EmptyAcquire counter показывает обращения к пустому пулу. Busy/capacity — текущая занятость обработчиков, а не CPU%. E2E время измерено от created_at job в транзакции durable acceptance до COMMIT ready, включает queue wait и retry. Replay использует исходное время принятия. + +Percentiles вычисляются объединением histogram buckets реплик: + +```promql +histogram_quantile(0.95, + sum by (le) (rate(messenger_http_duration_seconds_bucket{service_name="messenger-api",route!~"/health/.*"}[1m]))) +``` + +p50 и p99 используют 0.50 и 0.99. Единицы всех duration panels — секунды. Суммируются buckets, не percentiles отдельных реплик. Полные исполняемые запросы находятся в `grafana/dashboards/messenger.json`. + +## Логи и трассы + +API log event `http.request`: timestamp, severity, service.name, method, route template, status, duration_ms, request_id, trace_id, span_id. Worker — `job.started`, `job.completed`; dispatcher — `job.published`. + +Технические request_id/job_id допустимы только в логах/трассах для корреляции синтетических работ. В labels метрик они отсутствуют. SQL text/args, URL/query, filename, headers Authorization, пароли, токены, текст сообщения и содержимое файлов не записываются. + +HTTP server span принимает W3C traceparent/tracestate. Upload транзакционно сохраняет их в job. Producer span публикует свой контекст в разрешённых полях JSON брокера; consumer span извлекает его как parent. Повторы образуют новые producer/consumer ветки того же trace. S3, Redis sessions, DB acquire/query/commit и image.process имеют реальные дочерние spans. + +Log body — структурный JSON, дополнительно отправляемый OTLP Log SDK с trace/span context. В Grafana Loki TraceID derived field открывает Tempo; поиск по trace_id показывает связанные события. Заголовок X-Trace-Id добавлен для диагностики; X-Request-Id и бизнес-контракт сохранены. + +## Буферизация + +SDK: batch queue 256, batch 64, export interval 500 ms, exporter timeout 2 s, retry SDK выключен; metrics interval 2 s. Collector: memory limiter 128 MiB со spike 32 MiB, hard limit контейнера 256 MiB; batch size 64/max128; по 32 batches и 2 consumers для traces/logs, неблокирующее переполнение, retry deadline 5 s. Метрики предоставляются ограниченным набором series на pull endpoint Collector; Prometheus хранит историю в volume. Retention storage — 24 h. + +При отказе telemetry допускаются потери signals. Ошибки SDK учитываются отдельно; Collector показывает failed exports/queue, а финальные drops записывает с dropped_items. Business transaction от успешности экспорта не зависит. diff --git a/infra/otel/tempo.yaml b/infra/otel/tempo.yaml new file mode 100644 index 0000000..f0dc9f2 --- /dev/null +++ b/infra/otel/tempo.yaml @@ -0,0 +1,22 @@ +server: + http_listen_port: 3200 +distributor: + receivers: + otlp: + protocols: + http: + endpoint: 0.0.0.0:4318 + grpc: + endpoint: 0.0.0.0:4317 +ingester: + max_block_duration: 1m +compactor: + compaction: + block_retention: 24h +storage: + trace: + backend: local + wal: + path: /var/tempo/wal + local: + path: /var/tempo/blocks diff --git a/internal/database/migrate.go b/internal/database/migrate.go new file mode 100644 index 0000000..7304182 --- /dev/null +++ b/internal/database/migrate.go @@ -0,0 +1,64 @@ +package database + +import ( + "context" + "embed" + "fmt" + "io/fs" + "sort" + "strconv" + "strings" + + "github.com/jackc/pgx/v5" + "github.com/jackc/pgx/v5/pgxpool" +) + +//go:embed migrations/*.sql +var migrations embed.FS + +func Migrate(ctx context.Context, pool *pgxpool.Pool) error { + tx, err := pool.BeginTx(ctx, pgx.TxOptions{}) + if err != nil { + return err + } + defer tx.Rollback(ctx) + if _, err = tx.Exec(ctx, `SELECT pg_advisory_xact_lock(202602)`); err != nil { + return err + } + if _, err = tx.Exec(ctx, `CREATE TABLE IF NOT EXISTS schema_migrations (version integer PRIMARY KEY, applied_at timestamptz NOT NULL DEFAULT clock_timestamp())`); err != nil { + return err + } + entries, err := fs.ReadDir(migrations, "migrations") + if err != nil { + return err + } + sort.Slice(entries, func(i, j int) bool { return entries[i].Name() < entries[j].Name() }) + for _, entry := range entries { + if entry.IsDir() || !strings.HasSuffix(entry.Name(), ".sql") { + continue + } + part := strings.SplitN(entry.Name(), "_", 2)[0] + version, parseErr := strconv.Atoi(part) + if parseErr != nil { + return fmt.Errorf("invalid migration name %s", entry.Name()) + } + var applied bool + if err = tx.QueryRow(ctx, `SELECT EXISTS(SELECT 1 FROM schema_migrations WHERE version=$1)`, version).Scan(&applied); err != nil { + return err + } + if applied { + continue + } + sqlBytes, readErr := migrations.ReadFile("migrations/" + entry.Name()) + if readErr != nil { + return readErr + } + if _, err = tx.Exec(ctx, string(sqlBytes)); err != nil { + return fmt.Errorf("migration %s: %w", entry.Name(), err) + } + if _, err = tx.Exec(ctx, `INSERT INTO schema_migrations(version) VALUES($1)`, version); err != nil { + return err + } + } + return tx.Commit(ctx) +} diff --git a/internal/database/migrations/001_initial.sql b/internal/database/migrations/001_initial.sql new file mode 100644 index 0000000..72809a0 --- /dev/null +++ b/internal/database/migrations/001_initial.sql @@ -0,0 +1,31 @@ +CREATE TABLE users ( + id uuid PRIMARY KEY, + username text NOT NULL UNIQUE CHECK (username ~ '^[a-z0-9_]{3,32}$'), + display_name text NOT NULL CHECK (char_length(btrim(display_name)) > 0 AND char_length(display_name) <= 100), + password_hash text NOT NULL, + created_at timestamptz NOT NULL DEFAULT clock_timestamp() +); + +CREATE TABLE chats ( + id uuid PRIMARY KEY, + title text NOT NULL CHECK (char_length(btrim(title)) > 0 AND char_length(title) <= 100), + created_at timestamptz NOT NULL DEFAULT clock_timestamp() +); + +CREATE TABLE chat_members ( + chat_id uuid NOT NULL REFERENCES chats(id) ON DELETE CASCADE, + user_id uuid NOT NULL REFERENCES users(id) ON DELETE RESTRICT, + PRIMARY KEY (chat_id, user_id) +); + +CREATE INDEX chat_members_user_chat_idx ON chat_members (user_id, chat_id); + +CREATE TABLE messages ( + id uuid PRIMARY KEY, + chat_id uuid NOT NULL REFERENCES chats(id) ON DELETE CASCADE, + sender_id uuid NOT NULL REFERENCES users(id) ON DELETE RESTRICT, + text text NOT NULL CHECK (char_length(btrim(text)) > 0 AND char_length(text) <= 4000), + created_at timestamptz NOT NULL DEFAULT clock_timestamp() +); + +CREATE INDEX messages_chat_history_idx ON messages (chat_id, created_at, id); diff --git a/internal/database/migrations/002_message_idempotency.sql b/internal/database/migrations/002_message_idempotency.sql new file mode 100644 index 0000000..13b4d02 --- /dev/null +++ b/internal/database/migrations/002_message_idempotency.sql @@ -0,0 +1,13 @@ +CREATE TABLE message_idempotency ( + user_id uuid NOT NULL REFERENCES users(id) ON DELETE CASCADE, + method text NOT NULL, + path text NOT NULL, + idempotency_key text NOT NULL, + request_hash bytea NOT NULL, + message_id uuid REFERENCES messages(id) ON DELETE CASCADE, + created_at timestamptz NOT NULL DEFAULT clock_timestamp(), + PRIMARY KEY (user_id, method, path, idempotency_key) +); + +CREATE INDEX message_idempotency_created_at_idx + ON message_idempotency (created_at); diff --git a/internal/database/migrations/003_attachments.sql b/internal/database/migrations/003_attachments.sql new file mode 100644 index 0000000..af5a02f --- /dev/null +++ b/internal/database/migrations/003_attachments.sql @@ -0,0 +1,30 @@ +CREATE TABLE attachments ( + id uuid PRIMARY KEY, + chat_id uuid NOT NULL REFERENCES chats(id) ON DELETE CASCADE, + owner_id uuid NOT NULL REFERENCES users(id), + filename text NOT NULL, + object_key text NOT NULL UNIQUE, + content_type text NOT NULL CHECK (content_type IN ('image/png','image/jpeg')), + size_bytes bigint NOT NULL CHECK (size_bytes BETWEEN 1 AND 10485760), + sha256 text NOT NULL CHECK (sha256 ~ '^[a-f0-9]{64}$'), + status text NOT NULL DEFAULT 'ready' CHECK (status='ready'), + created_at timestamptz NOT NULL DEFAULT clock_timestamp(), + updated_at timestamptz NOT NULL DEFAULT clock_timestamp() +); +CREATE INDEX attachments_chat_idx ON attachments(chat_id,created_at,id); + +-- Durable intent is committed BEFORE any S3 PUT. A crashed upload remains +-- visible to the cleanup process even if its ready metadata never committed. +CREATE TABLE attachment_uploads ( + id uuid PRIMARY KEY, + object_key text NOT NULL UNIQUE, + created_at timestamptz NOT NULL DEFAULT clock_timestamp() +); +CREATE INDEX attachment_uploads_created_idx ON attachment_uploads(created_at); + +ALTER TABLE messages DROP CONSTRAINT messages_text_check; +ALTER TABLE messages ADD CONSTRAINT messages_text_check CHECK (char_length(text)<=4000); +ALTER TABLE messages ADD COLUMN attachment_ids uuid[] NOT NULL DEFAULT '{}'; +ALTER TABLE messages ADD CONSTRAINT messages_body_check + CHECK (char_length(btrim(text))>0 OR cardinality(attachment_ids)>0); +ALTER TABLE messages ADD CONSTRAINT messages_attachments_limit CHECK (cardinality(attachment_ids)<=10); diff --git a/internal/database/migrations/004_async_pipeline.sql b/internal/database/migrations/004_async_pipeline.sql new file mode 100644 index 0000000..16ae51e --- /dev/null +++ b/internal/database/migrations/004_async_pipeline.sql @@ -0,0 +1,37 @@ +ALTER TABLE attachments DROP CONSTRAINT attachments_status_check; +ALTER TABLE attachments ALTER COLUMN status SET DEFAULT 'queued'; +ALTER TABLE attachments ADD CONSTRAINT attachments_status_check CHECK (status IN ('queued','processing','ready','failed')); +ALTER TABLE attachments ADD COLUMN error_code text; +ALTER TABLE attachments ADD COLUMN variants jsonb NOT NULL DEFAULT '[]'; + +CREATE TABLE attachment_jobs ( + id uuid PRIMARY KEY, + attachment_id uuid NOT NULL UNIQUE REFERENCES attachments(id) ON DELETE CASCADE, + pipeline_version integer NOT NULL DEFAULT 1 CHECK (pipeline_version=1), + staging_key text NOT NULL UNIQUE, + state text NOT NULL DEFAULT 'queued' CHECK (state IN ('queued','processing','ready','failed')), + attempts integer NOT NULL DEFAULT 0, + max_attempts integer NOT NULL DEFAULT 5 CHECK (max_attempts BETWEEN 1 AND 20), + next_at timestamptz NOT NULL DEFAULT clock_timestamp(), + lease_until timestamptz, + lease_token uuid, + error_code text, + staging_deleted_at timestamptz, + created_at timestamptz NOT NULL DEFAULT clock_timestamp(), + updated_at timestamptz NOT NULL DEFAULT clock_timestamp() +); +CREATE INDEX attachment_jobs_due_idx ON attachment_jobs(state,next_at); +-- Same transaction as attachment + job; repeated publication is intentional. +CREATE TABLE attachment_outbox ( + job_id uuid PRIMARY KEY REFERENCES attachment_jobs(id) ON DELETE CASCADE, + published_at timestamptz, + publications integer NOT NULL DEFAULT 0 +); +CREATE TABLE attachment_replays ( + id bigserial PRIMARY KEY, + job_id uuid NOT NULL REFERENCES attachment_jobs(id), + previous_error text NOT NULL, + replayed_at timestamptz NOT NULL DEFAULT clock_timestamp() +); +-- Internal CLI-only, one-shot crash injection. No HTTP endpoint exposes it. +CREATE TABLE pipeline_faults (point text PRIMARY KEY, armed boolean NOT NULL DEFAULT true, hit_at timestamptz); diff --git a/internal/database/migrations/005_trace_context.sql b/internal/database/migrations/005_trace_context.sql new file mode 100644 index 0000000..bbbdc6f --- /dev/null +++ b/internal/database/migrations/005_trace_context.sql @@ -0,0 +1,4 @@ +-- Persist causal context in the same transaction as durable acceptance/outbox. +ALTER TABLE attachment_jobs ADD COLUMN traceparent text NOT NULL DEFAULT ''; +ALTER TABLE attachment_jobs ADD COLUMN tracestate text NOT NULL DEFAULT ''; +ALTER TABLE attachment_jobs ADD COLUMN request_id text NOT NULL DEFAULT ''; diff --git a/internal/messenger/attachments.go b/internal/messenger/attachments.go new file mode 100644 index 0000000..129c758 --- /dev/null +++ b/internal/messenger/attachments.go @@ -0,0 +1,263 @@ +package messenger + +import ( + "context" + "crypto/sha256" + "encoding/hex" + "encoding/json" + "errors" + "image" + _ "image/jpeg" + _ "image/png" + "io" + "mime" + "net/http" + "net/url" + "os" + "strconv" + "strings" + "time" + "unicode" + "unicode/utf8" + + "github.com/jackc/pgx/v5" + "mtusi-messenger/internal/telemetry" +) + +const MaxUploadBytes int64 = 10 * 1024 * 1024 +const MaxImagePixels int64 = 40000000 + +type Attachment struct { + ID string `json:"id"` + ChatID string `json:"chat_id"` + OwnerID string `json:"owner_id"` + Filename string `json:"filename"` + ContentType string `json:"content_type"` + SizeBytes int64 `json:"size_bytes"` + SHA256 string `json:"sha256"` + Status string `json:"status"` + Variants []any `json:"variants"` + ErrorCode *string `json:"error_code"` + CreatedAt string `json:"created_at"` + UpdatedAt string `json:"updated_at"` +} + +func tooLarge() *apiError { + return &apiError{status: 413, code: "payload_too_large", message: "Превышен лимит изображения"} +} +func invalidImage() *apiError { + return &apiError{status: 415, code: "unsupported_media_type", message: "Требуется корректный PNG или JPEG"} +} + +func rollbackUpload(tx pgx.Tx) { + ctx, cancel := context.WithTimeout(context.Background(), 3*time.Second) + defer cancel() + _ = tx.Rollback(ctx) +} + +func validFilename(s string) bool { + if !utf8.ValidString(s) || strings.TrimSpace(s) == "" || utf8.RuneCountInString(s) > 255 { + return false + } + for _, c := range s { + if c == '/' || c == '\\' || unicode.IsControl(c) { + return false + } + } + return true +} + +func (a *API) isMember(r *http.Request, uid, chatID string) *apiError { + ctx, cancel := a.dbContext(r.Context()) + defer cancel() + var member bool + if err := a.db.QueryRow(ctx, `SELECT EXISTS(SELECT 1 FROM chat_members WHERE chat_id=$1 AND user_id=$2)`, chatID, uid).Scan(&member); err != nil { + return unavailable() + } + if !member { + return missing() + } + return nil +} + +func (a *API) uploadAttachment(r *http.Request, uid, chatID string) (response, *apiError) { + if e := a.isMember(r, uid, chatID); e != nil { + return response{}, e + } + query, err := url.ParseQuery(r.URL.RawQuery) + if err != nil || len(query) != 1 || len(query["filename"]) != 1 || !validFilename(query.Get("filename")) { + return response{}, bad("Неверное имя файла") + } + kind, _, err := mime.ParseMediaType(r.Header.Get("Content-Type")) + if err != nil || (kind != "image/png" && kind != "image/jpeg") { + return response{}, invalidImage() + } + if r.ContentLength > MaxUploadBytes { + return response{}, tooLarge() + } + ctx, cancel := context.WithTimeout(r.Context(), 45*time.Second) + defer cancel() + // One decoder/temporary original per replica bounds RAM and tmpfs usage. + select { + case a.uploadSlots <- struct{}{}: + defer func() { <-a.uploadSlots }() + case <-ctx.Done(): + return response{}, unavailable() + } + file, err := os.CreateTemp("", "upload-*") + if err != nil { + return response{}, unavailable() + } + defer os.Remove(file.Name()) + defer file.Close() + hash := sha256.New() + size, err := io.CopyBuffer(io.MultiWriter(file, hash), io.LimitReader(r.Body, MaxUploadBytes+1), make([]byte, 32*1024)) + if size > MaxUploadBytes { + return response{}, tooLarge() + } + if err != nil { + return response{}, bad("Загрузка прервана") + } + if size == 0 { + return response{}, invalidImage() + } + file.Seek(0, io.SeekStart) + config, format, err := image.DecodeConfig(file) + if err != nil || (format != "png" && format != "jpeg") || "image/"+format != kind { + return response{}, invalidImage() + } + if config.Width < 1 || config.Height < 1 || int64(config.Width) > MaxImagePixels/int64(config.Height) { + return response{}, tooLarge() + } + file.Seek(0, io.SeekStart) + // Decode after the dimension guard to reject corrupt payloads, not just + // apparently valid headers. The decoded bitmap is never persisted. + if e := validateImage(file); e != nil { + return response{}, e + } + file.Seek(0, io.SeekStart) + id := newID() + key := "staging/" + id + dbctx, dbcancel := a.dbContext(ctx) + _, err = a.db.Exec(dbctx, `INSERT INTO attachment_uploads(id,object_key) VALUES($1,$2)`, id, key) + dbcancel() + if err != nil { + return response{}, unavailable() + } + tx, err := a.db.Begin(ctx) + if err != nil { + return response{}, unavailable() + } + defer rollbackUpload(tx) + // Cleanup and upload use the same lock. Cleanup cannot delete an object + // while its upload/metadata transaction is in progress. + if _, err = tx.Exec(ctx, `SELECT pg_advisory_xact_lock(hashtextextended($1,4))`, key); err != nil { + return response{}, unavailable() + } + var pending bool + if err = tx.QueryRow(ctx, `SELECT EXISTS(SELECT 1 FROM attachment_uploads WHERE id=$1)`, id).Scan(&pending); err != nil || !pending { + return response{}, unavailable() + } + putctx, putcancel := context.WithTimeout(ctx, 20*time.Second) + err = a.objects.Put(putctx, key, file, size, kind) + putcancel() + if err != nil { + return response{}, unavailable() + } + att := Attachment{ID: id, ChatID: chatID, OwnerID: uid, Filename: query.Get("filename"), ContentType: kind, SizeBytes: size, SHA256: hex.EncodeToString(hash.Sum(nil)), Status: "queued", Variants: []any{}} + var created, updated time.Time + if err = tx.QueryRow(ctx, `INSERT INTO attachments(id,chat_id,owner_id,filename,object_key,content_type,size_bytes,sha256,status) VALUES($1,$2,$3,$4,$5,$6,$7,$8,'queued') RETURNING created_at,updated_at`, id, chatID, uid, att.Filename, key, kind, size, att.SHA256).Scan(&created, &updated); err != nil { + return response{}, unavailable() + } + if _, err = tx.Exec(ctx, `DELETE FROM attachment_uploads WHERE id=$1`, id); err != nil { + return response{}, unavailable() + } + + maxAttempts := 5 + if raw := os.Getenv("JOB_MAX_ATTEMPTS"); raw != "" { + n, parseErr := strconv.Atoi(raw) + if parseErr != nil || n < 1 || n > 20 { + return response{}, unavailable() + } + maxAttempts = n + } + parent, state := telemetry.Inject(ctx) + if _, err = tx.Exec(ctx, `INSERT INTO attachment_jobs(id,attachment_id,staging_key,max_attempts,traceparent,tracestate,request_id) VALUES($1,$2,$3,$4,$5,$6,$7)`, id, id, key, maxAttempts, parent, state, telemetry.RequestID(ctx)); err != nil { + return response{}, unavailable() + } + if _, err = tx.Exec(ctx, `INSERT INTO attachment_outbox(job_id) VALUES($1)`, id); err != nil { + return response{}, unavailable() + } + if err = tx.Commit(ctx); err != nil { + return response{}, unavailable() + } + att.CreatedAt = formatTime(created) + att.UpdatedAt = formatTime(updated) + return response{status: 202, body: att}, nil +} + +func validateImage(r io.Reader) *apiError { + _, _, err := image.Decode(r) + if err != nil { + return invalidImage() + } + return nil +} + +func (a *API) getAttachment(r *http.Request, uid, id string, content bool) (response, *apiError) { + query, err := url.ParseQuery(r.URL.RawQuery) + if err != nil { + return response{}, bad("Неверные параметры вложения") + } + if (!content && len(query) > 0) || (content && (len(query) > 1 || (len(query) > 0 && len(query["variant"]) != 1))) { + return response{}, bad("Неверные параметры вложения") + } + ctx, cancel := a.dbContext(r.Context()) + defer cancel() + att := Attachment{Variants: []any{}} + var key string + var variantsJSON []byte + var created, updated time.Time + err = a.db.QueryRow(ctx, `SELECT a.id::text,a.chat_id::text,a.owner_id::text,a.filename,a.object_key,a.content_type,a.size_bytes,a.sha256,a.status,a.created_at,a.updated_at,a.variants,a.error_code FROM attachments a JOIN chat_members cm ON cm.chat_id=a.chat_id AND cm.user_id=$2 WHERE a.id=$1`, id, uid).Scan(&att.ID, &att.ChatID, &att.OwnerID, &att.Filename, &key, &att.ContentType, &att.SizeBytes, &att.SHA256, &att.Status, &created, &updated, &variantsJSON, &att.ErrorCode) + if errors.Is(err, pgx.ErrNoRows) { + return response{}, missing() + } + if err != nil { + return response{}, unavailable() + } + _ = json.Unmarshal(variantsJSON, &att.Variants) + att.CreatedAt = formatTime(created) + att.UpdatedAt = formatTime(updated) + if !content { + return response{status: 200, body: att}, nil + } + + if att.Status != "ready" { + return response{}, &apiError{status: 409, code: "conflict", message: "Вложение не готово"} + } + if variant, ok := query["variant"]; ok && variant[0] != "original" { + found := false + for _, item := range att.Variants { + v, ok := item.(map[string]any) + if ok && v["name"] == variant[0] { + key = "variants/" + id + "/v1/" + variant[0] + att.ContentType = v["content_type"].(string) + att.SizeBytes = int64(v["size_bytes"].(float64)) + att.Filename = variant[0] + map[bool]string{true: ".webp", false: ".jpg"}[variant[0] == "optimized"] + found = true + break + } + } + if !found { + return response{}, missing() + } + } + + downloadCtx, downloadCancel := context.WithTimeout(r.Context(), 45*time.Second) + object, err := a.objects.Get(downloadCtx, key) + if err != nil { + downloadCancel() + return response{}, unavailable() + } + return response{status: 200, content: object, contentType: att.ContentType, length: att.SizeBytes, filename: att.Filename, cancel: downloadCancel}, nil +} diff --git a/internal/messenger/cleanup.go b/internal/messenger/cleanup.go new file mode 100644 index 0000000..797edb0 --- /dev/null +++ b/internal/messenger/cleanup.go @@ -0,0 +1,118 @@ +package messenger + +import ( + "context" + "fmt" + "strings" + "time" + + "github.com/jackc/pgx/v5/pgxpool" + "github.com/minio/minio-go/v7" + "mtusi-messenger/internal/objectstore" +) + +type CleanupResult struct { + RemovedObjects int `json:"removed_objects"` + RemovedIntents int64 `json:"removed_intents"` + KeptReady int `json:"kept_ready"` + SkippedRecent int `json:"skipped_recent"` + SkippedBusy int `json:"skipped_busy"` +} + +// Cleanup removes abandoned uploads. Accepted pipeline jobs protect staging. +func Cleanup(ctx context.Context, db *pgxpool.Pool, store *objectstore.Store, grace time.Duration) (CleanupResult, error) { + result := CleanupResult{} + cutoff := time.Now().Add(-grace) + clean := func(key string) error { + if !strings.HasPrefix(key, "originals/") && !strings.HasPrefix(key, "staging/") { + return nil + } + tx, err := db.Begin(ctx) + if err != nil { + return err + } + defer rollbackUpload(tx) + var locked bool + if err = tx.QueryRow(ctx, `SELECT pg_try_advisory_xact_lock(hashtextextended($1,4))`, key).Scan(&locked); err != nil { + return err + } + if !locked { + result.SkippedBusy++ + return nil + } + var ready, recent bool + if err = tx.QueryRow(ctx, `SELECT (EXISTS(SELECT 1 FROM attachments WHERE object_key=$1) OR EXISTS(SELECT 1 FROM attachment_jobs WHERE staging_key=$1)),EXISTS(SELECT 1 FROM attachment_uploads WHERE object_key=$1 AND created_at>$2)`, key, cutoff).Scan(&ready, &recent); err != nil { + return err + } + if ready { + result.KeptReady++ + return nil + } + if recent { + result.SkippedRecent++ + return nil + } + info, err := store.Client.StatObject(ctx, store.Bucket, key, minio.StatObjectOptions{}) + exists := err == nil + if err != nil && minio.ToErrorResponse(err).Code != "NoSuchKey" { + return err + } + if exists && info.LastModified.After(cutoff) { + result.SkippedRecent++ + return nil + } + if exists { + if err = store.Delete(ctx, key); err != nil { + return err + } + } + command, err := tx.Exec(ctx, `DELETE FROM attachment_uploads WHERE object_key=$1`, key) + if err != nil { + return err + } + if err = tx.Commit(ctx); err != nil { + return err + } + result.RemovedIntents += command.RowsAffected() + if exists { + result.RemovedObjects++ + } + return nil + } + rows, err := db.Query(ctx, `SELECT object_key FROM attachment_uploads WHERE created_at<$1 ORDER BY created_at LIMIT 1000`, cutoff) + if err != nil { + return result, err + } + keys := []string{} + for rows.Next() { + var key string + if err = rows.Scan(&key); err != nil { + rows.Close() + return result, err + } + keys = append(keys, key) + } + err = rows.Err() + rows.Close() + if err != nil { + return result, err + } + for _, key := range keys { + if err = clean(key); err != nil { + return result, err + } + } + listctx, cancel := context.WithCancel(ctx) + defer cancel() + for object := range store.Client.ListObjects(listctx, store.Bucket, minio.ListObjectsOptions{Recursive: true}) { + if object.Err != nil { + return result, fmt.Errorf("S3 cleanup listing: %w", object.Err) + } + if object.LastModified.Before(cutoff) { + if err = clean(object.Key); err != nil { + return result, err + } + } + } + return result, nil +} diff --git a/internal/messenger/password.go b/internal/messenger/password.go new file mode 100644 index 0000000..4b0d0ec --- /dev/null +++ b/internal/messenger/password.go @@ -0,0 +1,44 @@ +package messenger + +import ( + "crypto/rand" + "crypto/subtle" + "encoding/base64" + "fmt" + "strconv" + "strings" + + "golang.org/x/crypto/scrypt" +) + +const scryptN, scryptR, scryptP = 32768, 8, 1 + +func hashPassword(password string) (string, error) { + salt := make([]byte, 16) + if _, err := rand.Read(salt); err != nil { + return "", err + } + hash, err := scrypt.Key([]byte(password), salt, scryptN, scryptR, scryptP, 32) + if err != nil { + return "", err + } + b64 := base64.RawStdEncoding + return fmt.Sprintf("scrypt$%d$%d$%d$%s$%s", scryptN, scryptR, scryptP, b64.EncodeToString(salt), b64.EncodeToString(hash)), nil +} + +func verifyPassword(encoded, password string) bool { + parts := strings.Split(encoded, "$") + if len(parts) != 6 || parts[0] != "scrypt" { + return false + } + n, e1 := strconv.Atoi(parts[1]) + r, e2 := strconv.Atoi(parts[2]) + p, e3 := strconv.Atoi(parts[3]) + salt, e4 := base64.RawStdEncoding.DecodeString(parts[4]) + expected, e5 := base64.RawStdEncoding.DecodeString(parts[5]) + if e1 != nil || e2 != nil || e3 != nil || e4 != nil || e5 != nil || len(expected) != 32 { + return false + } + actual, err := scrypt.Key([]byte(password), salt, n, r, p, len(expected)) + return err == nil && subtle.ConstantTimeCompare(actual, expected) == 1 +} diff --git a/internal/messenger/password_test.go b/internal/messenger/password_test.go new file mode 100644 index 0000000..bf05e6e --- /dev/null +++ b/internal/messenger/password_test.go @@ -0,0 +1,33 @@ +package messenger + +import ( + "strings" + "testing" +) + +func TestPasswordHash(t *testing.T) { + const password = "long-test-password" + first, err := hashPassword(password) + if err != nil { + t.Fatal(err) + } + second, err := hashPassword(password) + if err != nil { + t.Fatal(err) + } + if first == password || strings.Contains(first, password) { + t.Fatal("hash contains the original password") + } + if first == second { + t.Fatal("independent hashes must use different salts") + } + if !verifyPassword(first, password) { + t.Fatal("correct password was rejected") + } + if verifyPassword(first, "wrong-test-password") { + t.Fatal("wrong password was accepted") + } + if verifyPassword("damaged-hash", password) { + t.Fatal("malformed hash was accepted") + } +} diff --git a/internal/messenger/server.go b/internal/messenger/server.go new file mode 100644 index 0000000..41587b1 --- /dev/null +++ b/internal/messenger/server.go @@ -0,0 +1,776 @@ +package messenger + +import ( + "bytes" + "context" + "crypto/rand" + "crypto/sha256" + "encoding/base64" + "encoding/hex" + "encoding/json" + "errors" + "fmt" + "io" + "log" + "mime" + "net/http" + "net/url" + "regexp" + "strconv" + "strings" + "time" + "unicode/utf8" + + "github.com/jackc/pgx/v5" + "github.com/jackc/pgx/v5/pgconn" + "github.com/jackc/pgx/v5/pgxpool" + "github.com/redis/go-redis/v9" + "go.opentelemetry.io/otel" + "go.opentelemetry.io/otel/attribute" + "go.opentelemetry.io/otel/codes" + "go.opentelemetry.io/otel/propagation" + "go.opentelemetry.io/otel/trace" + "mtusi-messenger/internal/objectstore" + "mtusi-messenger/internal/telemetry" +) + +type User struct { + ID string `json:"id"` + Username string `json:"username"` + DisplayName string `json:"display_name"` + CreatedAt string `json:"created_at"` +} +type Chat struct { + ID string `json:"id"` + Title string `json:"title"` + MemberIDs []string `json:"member_ids"` + CreatedAt string `json:"created_at"` +} +type Message struct { + ID string `json:"id"` + ChatID string `json:"chat_id"` + SenderID string `json:"sender_id"` + Text string `json:"text"` + CreatedAt string `json:"created_at"` + AttachmentIDs []string `json:"attachment_ids"` +} +type Session struct { + Token string `json:"token"` + TokenType string `json:"token_type"` + ExpiresAt string `json:"expires_at"` + User User `json:"user"` +} + +type API struct { + objects *objectstore.Store + uploadSlots chan struct{} + db *pgxpool.Pool + sessions *redis.Client + ttl time.Duration + idempotencyTTL time.Duration + dbTimeout time.Duration + instanceID string + logger *log.Logger +} + +func NewHandler(db *pgxpool.Pool, sessions *redis.Client, objects *objectstore.Store, ttl, idempotencyTTL, dbTimeout time.Duration, instanceID string, logger *log.Logger) http.Handler { + return &API{db: db, sessions: sessions, objects: objects, uploadSlots: make(chan struct{}, 1), ttl: ttl, idempotencyTTL: idempotencyTTL, dbTimeout: dbTimeout, instanceID: instanceID, logger: logger} +} + +var uuidRE = regexp.MustCompile(`^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$`) +var usernameRE = regexp.MustCompile(`^[a-z0-9_]{3,32}$`) +var idempotencyKeyRE = regexp.MustCompile(`^[A-Za-z0-9._:-]{1,128}$`) + +type apiError struct { + status int + code, message, challenge string +} +type response struct { + status int + body any + noBody bool + content io.ReadCloser + contentType string + length int64 + filename string + cancel context.CancelFunc +} + +func bad(msg string) *apiError { return &apiError{status: 400, code: "invalid_request", message: msg} } +func missing() *apiError { + return &apiError{status: 404, code: "not_found", message: "Объект не найден"} +} +func unavailable() *apiError { + return &apiError{status: 503, code: "unavailable", message: "Сервис временно недоступен"} +} +func unauthorized(scheme string) *apiError { + challenge := `Bearer realm="messenger"` + if scheme == "Basic" { + challenge = `Basic realm="messenger", charset="UTF-8"` + } + return &apiError{status: 401, code: "unauthorized", message: "Неверные учётные данные", challenge: challenge} +} + +func (a *API) ServeHTTP(w http.ResponseWriter, r *http.Request) { + + id := newID() + route, method := telemetry.Route(r.URL.Path), telemetry.Method(r.Method) + parent := otel.GetTextMapPropagator().Extract(r.Context(), propagation.HeaderCarrier(r.Header)) + ctx, span := telemetry.Span(parent, method+" "+route, trace.SpanKindServer, attribute.String("http.route", route), attribute.String("http.request.method", method), attribute.String("request.id", id)) + ctx = telemetry.WithRequestID(ctx, id) + r = r.WithContext(ctx) + w.Header().Set("X-Trace-Id", span.SpanContext().TraceID().String()) + w.Header().Set("X-Request-Id", id) + w.Header().Set("X-Instance-Id", a.instanceID) + w.Header().Set("Content-Type", "application/json; charset=utf-8") + start, status := time.Now(), 500 + defer func() { + if x := recover(); x != nil { + status = 500 + telemetry.Event(ctx, "http.panic", nil) + writeError(w, id, &apiError{status: 500, code: "internal_error", message: "Внутренняя ошибка"}) + } + span.SetAttributes(attribute.Int("http.response.status_code", status)) + if status >= 500 { + span.SetStatus(codes.Error, "HTTP server error") + } + telemetry.HTTP(ctx, method, route, status, time.Since(start)) + telemetry.Event(ctx, "http.request", map[string]any{"method": method, "route": route, "status": status, "duration_ms": time.Since(start).Milliseconds()}) + span.End() + }() + res, apiErr := a.route(r) + if apiErr != nil { + status = apiErr.status + writeError(w, id, apiErr) + return + } + status = res.status + if res.content != nil { + if res.cancel != nil { + defer res.cancel() + } + defer res.content.Close() + w.Header().Set("Content-Type", res.contentType) + w.Header().Set("Content-Length", strconv.FormatInt(res.length, 10)) + w.Header().Set("Content-Disposition", mime.FormatMediaType("inline", map[string]string{"filename": res.filename})) + w.Header().Set("Cache-Control", "private, no-store") + w.Header().Set("X-Content-Type-Options", "nosniff") + w.WriteHeader(status) + if _, err := io.Copy(w, res.content); err != nil { + telemetry.Event(ctx, "http.download_interrupted", nil) + } + return + } + w.WriteHeader(status) + if !res.noBody { + _ = json.NewEncoder(w).Encode(res.body) + } +} + +func writeError(w http.ResponseWriter, requestID string, e *apiError) { + if e.challenge != "" { + w.Header().Set("WWW-Authenticate", e.challenge) + } + w.WriteHeader(e.status) + _ = json.NewEncoder(w).Encode(map[string]any{"error": map[string]string{"code": e.code, "message": e.message, "request_id": requestID}}) +} + +func decode(r *http.Request, dst any) *apiError { + if strings.ToLower(strings.TrimSpace(strings.Split(r.Header.Get("Content-Type"), ";")[0])) != "application/json" { + return &apiError{status: 415, code: "unsupported_media_type", message: "Требуется application/json"} + } + if r.ContentLength > 1<<20 { + return &apiError{status: 413, code: "payload_too_large", message: "Слишком большой запрос"} + } + data, err := io.ReadAll(io.LimitReader(r.Body, (1<<20)+1)) + if err != nil { + return bad("Не удалось прочитать JSON") + } + if len(data) > 1<<20 { + return &apiError{status: 413, code: "payload_too_large", message: "Слишком большой запрос"} + } + if !utf8.Valid(data) { + return bad("JSON должен быть в UTF-8") + } + var fields map[string]json.RawMessage + if json.Unmarshal(data, &fields) == nil { + for _, value := range fields { + if bytes.Equal(bytes.TrimSpace(value), []byte("null")) { + return bad("Поля запроса не должны быть null") + } + } + } + d := json.NewDecoder(bytes.NewReader(data)) + d.DisallowUnknownFields() + if err = d.Decode(dst); err != nil { + return bad("Неверный JSON или поля") + } + var rest any + if !errors.Is(d.Decode(&rest), io.EOF) { + return bad("Ожидался один JSON-объект") + } + return nil +} + +func validText(s string, max int) bool { + return utf8.ValidString(s) && strings.TrimSpace(s) != "" && utf8.RuneCountInString(s) <= max +} +func validPassword(s string) bool { + n := utf8.RuneCountInString(s) + return utf8.ValidString(s) && n >= 12 && n <= 128 +} +func formatTime(t time.Time) string { return t.UTC().Format(time.RFC3339Nano) } +func newID() string { + var b [16]byte + if _, err := rand.Read(b[:]); err != nil { + panic(err) + } + b[6] = (b[6] & 0x0f) | 0x40 + b[8] = (b[8] & 0x3f) | 0x80 + return fmt.Sprintf("%x-%x-%x-%x-%x", b[:4], b[4:6], b[6:8], b[8:10], b[10:]) +} +func newToken() string { + b := make([]byte, 32) + if _, err := rand.Read(b); err != nil { + panic(err) + } + return base64.RawURLEncoding.EncodeToString(b) +} +func tokenKey(token string) string { + sum := sha256.Sum256([]byte(token)) + return "session:" + hex.EncodeToString(sum[:]) +} + +func cursorFor(resource string, created time.Time, id string) string { + b, _ := json.Marshal([]string{resource, formatTime(created), id}) + return base64.RawURLEncoding.EncodeToString(b) +} +func parsePage(r *http.Request, resource string) (int, *time.Time, string, *apiError) { + q, err := url.ParseQuery(r.URL.RawQuery) + if err != nil { + return 0, nil, "", bad("Неверные параметры страницы") + } + for k, v := range q { + if (k != "limit" && k != "cursor") || len(v) != 1 { + return 0, nil, "", bad("Неверные параметры страницы") + } + } + limit := 50 + if raw := q.Get("limit"); raw != "" { + n, e := strconv.Atoi(raw) + if e != nil || n < 1 || n > 100 { + return 0, nil, "", bad("limit должен быть от 1 до 100") + } + limit = n + } else if _, ok := q["limit"]; ok { + return 0, nil, "", bad("Пустой limit") + } + if _, ok := q["cursor"]; !ok { + return limit, nil, "", nil + } + raw := q.Get("cursor") + if raw == "" || len(raw) > 1024 { + return 0, nil, "", bad("Неверный cursor") + } + b, e := base64.RawURLEncoding.DecodeString(raw) + if e != nil { + return 0, nil, "", bad("Неверный cursor") + } + var p []string + if json.Unmarshal(b, &p) != nil || len(p) != 3 || p[0] != resource || !uuidRE.MatchString(p[2]) { + return 0, nil, "", bad("Неверный cursor") + } + t, e := time.Parse(time.RFC3339Nano, p[1]) + if e != nil { + return 0, nil, "", bad("Неверный cursor") + } + return limit, &t, p[2], nil +} +func (a *API) dbContext(parent context.Context) (context.Context, context.CancelFunc) { + return context.WithTimeout(parent, a.dbTimeout) +} + +func (a *API) authenticate(r *http.Request) (string, string, *apiError) { + h := r.Header.Get("Authorization") + if !strings.HasPrefix(h, "Bearer ") || strings.TrimSpace(strings.TrimPrefix(h, "Bearer ")) == "" { + return "", "", unauthorized("Bearer") + } + token := strings.TrimSpace(strings.TrimPrefix(h, "Bearer ")) + ctx, cancel := context.WithTimeout(r.Context(), a.dbTimeout) + defer cancel() + uid, err := a.sessions.Get(ctx, tokenKey(token)).Result() + if errors.Is(err, redis.Nil) { + return "", "", unauthorized("Bearer") + } + if err != nil { + a.logger.Printf("session store unavailable: %v", err) + return "", "", unavailable() + } + if !uuidRE.MatchString(uid) { + return "", "", unauthorized("Bearer") + } + return uid, token, nil +} + +func scanUser(row pgx.Row) (User, string, error) { + var u User + var created time.Time + var password string + err := row.Scan(&u.ID, &u.Username, &u.DisplayName, &created, &password) + u.CreatedAt = formatTime(created) + return u, password, err +} + +func (a *API) route(r *http.Request) (response, *apiError) { + p := r.URL.Path + if r.Method == "GET" && p == "/health/live" { + return response{status: 200, body: map[string]string{"status": "ok"}}, nil + } + if r.Method == "GET" && p == "/health/ready" { + ctx, cancel := context.WithTimeout(r.Context(), time.Second) + defer cancel() + dbctx, dbspan := telemetry.Span(ctx, "db.readiness", trace.SpanKindClient, attribute.String("db.system.name", "postgresql")) + dbErr := a.db.Ping(dbctx) + telemetry.Result(dbspan, dbErr) + if dbErr != nil || a.sessions.Ping(ctx).Err() != nil || a.objects.Ready(ctx) != nil { + return response{}, unavailable() + } + return response{status: 200, body: map[string]string{"status": "ok"}}, nil + } + if r.Method == "POST" && p == "/api/v1/users" { + return a.createUser(r) + } + if r.Method == "POST" && p == "/api/v1/auth/sessions" { + return a.createSession(r) + } + if !strings.HasPrefix(p, "/api/v1/") { + return response{}, missing() + } + uid, token, authErr := a.authenticate(r) + if authErr != nil { + return response{}, authErr + } + if r.Method == "DELETE" && p == "/api/v1/auth/sessions/current" { + ctx, cancel := context.WithTimeout(r.Context(), a.dbTimeout) + defer cancel() + if err := a.sessions.Del(ctx, tokenKey(token)).Err(); err != nil { + return response{}, unavailable() + } + return response{status: 204, noBody: true}, nil + } + if r.Method == "GET" && p == "/api/v1/users/me" { + return a.currentUser(r, uid) + } + if strings.HasPrefix(p, "/api/v1/attachments/") { + parts := strings.Split(strings.TrimPrefix(p, "/api/v1/attachments/"), "/") + if !uuidRE.MatchString(parts[0]) { + return response{}, bad("Неверный attachment_id") + } + if r.Method == "GET" && (len(parts) == 1 || (len(parts) == 2 && parts[1] == "content")) { + return a.getAttachment(r, uid, parts[0], len(parts) == 2) + } + return response{}, missing() + } + if p == "/api/v1/chats" { + if r.Method == "POST" { + return a.createChat(r, uid) + } + if r.Method == "GET" { + return a.listChats(r, uid) + } + } + if !strings.HasPrefix(p, "/api/v1/chats/") { + return response{}, missing() + } + parts := strings.Split(strings.TrimPrefix(p, "/api/v1/chats/"), "/") + if len(parts) < 1 || !uuidRE.MatchString(parts[0]) { + return response{}, bad("Неверный chat_id") + } + chatID := parts[0] + if len(parts) == 2 && parts[1] == "attachments" && r.Method == "POST" { + return a.uploadAttachment(r, uid, chatID) + } + if len(parts) == 1 && r.Method == "GET" { + return a.getChat(r, uid, chatID) + } + if len(parts) == 2 && parts[1] == "messages" { + if r.Method == "POST" { + return a.createMessage(r, uid, chatID) + } + if r.Method == "GET" { + return a.listMessages(r, uid, chatID) + } + } + return response{}, missing() +} + +func (a *API) createUser(r *http.Request) (response, *apiError) { + var in struct { + Username string `json:"username"` + DisplayName string `json:"display_name"` + Password string `json:"password"` + } + if e := decode(r, &in); e != nil { + return response{}, e + } + if !usernameRE.MatchString(in.Username) || !validText(in.DisplayName, 100) || !validPassword(in.Password) { + return response{}, bad("Неверные данные пользователя") + } + passwordHash, err := hashPassword(in.Password) + if err != nil { + return response{}, &apiError{status: 500, code: "internal_error", message: "Внутренняя ошибка"} + } + ctx, cancel := a.dbContext(r.Context()) + defer cancel() + var u User + var created time.Time + err = a.db.QueryRow(ctx, `INSERT INTO users(id,username,display_name,password_hash) VALUES($1,$2,$3,$4) RETURNING id::text,username,display_name,created_at`, newID(), in.Username, in.DisplayName, passwordHash).Scan(&u.ID, &u.Username, &u.DisplayName, &created) + if err != nil { + var pgErr *pgconn.PgError + if errors.As(err, &pgErr) && pgErr.Code == "23505" { + return response{}, &apiError{status: 409, code: "conflict", message: "Имя пользователя занято"} + } + return response{}, unavailable() + } + u.CreatedAt = formatTime(created) + return response{status: 201, body: u}, nil +} + +func (a *API) createSession(r *http.Request) (response, *apiError) { + if r.ContentLength > 0 { + return response{}, bad("Тело запроса не допускается") + } + h := r.Header.Get("Authorization") + if !strings.HasPrefix(h, "Basic ") { + return response{}, unauthorized("Basic") + } + raw, err := base64.StdEncoding.DecodeString(strings.TrimSpace(strings.TrimPrefix(h, "Basic "))) + if err != nil { + return response{}, unauthorized("Basic") + } + parts := strings.SplitN(string(raw), ":", 2) + if len(parts) != 2 || !usernameRE.MatchString(parts[0]) { + return response{}, unauthorized("Basic") + } + ctx, cancel := a.dbContext(r.Context()) + defer cancel() + u, passwordHash, err := scanUser(a.db.QueryRow(ctx, `SELECT id::text,username,display_name,created_at,password_hash FROM users WHERE username=$1`, parts[0])) + if errors.Is(err, pgx.ErrNoRows) { + return response{}, unauthorized("Basic") + } + if err != nil { + return response{}, unavailable() + } + if !verifyPassword(passwordHash, parts[1]) { + return response{}, unauthorized("Basic") + } + token := newToken() + expires := time.Now().UTC().Add(a.ttl) + if err = a.sessions.Set(ctx, tokenKey(token), u.ID, a.ttl).Err(); err != nil { + return response{}, unavailable() + } + return response{status: 201, body: Session{Token: token, TokenType: "Bearer", ExpiresAt: formatTime(expires), User: u}}, nil +} + +func (a *API) currentUser(r *http.Request, uid string) (response, *apiError) { + ctx, cancel := a.dbContext(r.Context()) + defer cancel() + u, _, err := scanUser(a.db.QueryRow(ctx, `SELECT id::text,username,display_name,created_at,password_hash FROM users WHERE id=$1`, uid)) + if errors.Is(err, pgx.ErrNoRows) { + return response{}, unauthorized("Bearer") + } + if err != nil { + return response{}, unavailable() + } + return response{status: 200, body: u}, nil +} + +func (a *API) createChat(r *http.Request, uid string) (response, *apiError) { + var in struct { + Title string `json:"title"` + MemberIDs []string `json:"member_ids"` + } + if e := decode(r, &in); e != nil { + return response{}, e + } + if !validText(in.Title, 100) || len(in.MemberIDs) < 1 || len(in.MemberIDs) > 100 { + return response{}, bad("Неверные данные чата") + } + seen := map[string]bool{} + for _, id := range in.MemberIDs { + if !uuidRE.MatchString(id) || seen[id] { + return response{}, bad("Неверные участники чата") + } + seen[id] = true + } + if !seen[uid] { + return response{}, bad("Создатель должен быть участником") + } + ctx, cancel := a.dbContext(r.Context()) + defer cancel() + tx, err := a.db.Begin(ctx) + if err != nil { + return response{}, unavailable() + } + defer tx.Rollback(ctx) + c := Chat{ID: newID(), Title: in.Title, MemberIDs: append([]string{}, in.MemberIDs...)} + var created time.Time + if err = tx.QueryRow(ctx, `INSERT INTO chats(id,title) VALUES($1,$2) RETURNING created_at`, c.ID, c.Title).Scan(&created); err != nil { + return response{}, unavailable() + } + for _, id := range c.MemberIDs { + if _, err = tx.Exec(ctx, `INSERT INTO chat_members(chat_id,user_id) VALUES($1,$2)`, c.ID, id); err != nil { + var pgErr *pgconn.PgError + if errors.As(err, &pgErr) && pgErr.Code == "23503" { + return response{}, bad("Неверные участники чата") + } + return response{}, unavailable() + } + } + if err = tx.Commit(ctx); err != nil { + return response{}, unavailable() + } + c.CreatedAt = formatTime(created) + return response{status: 201, body: c}, nil +} + +func scanChat(row pgx.Row) (Chat, time.Time, error) { + var c Chat + var t time.Time + err := row.Scan(&c.ID, &c.Title, &c.MemberIDs, &t) + c.CreatedAt = formatTime(t) + return c, t, err +} +func chatSelect() string { + return `SELECT c.id::text,c.title,ARRAY(SELECT cm2.user_id::text FROM chat_members cm2 WHERE cm2.chat_id=c.id ORDER BY cm2.user_id),c.created_at FROM chats c` +} + +func (a *API) getChat(r *http.Request, uid, chatID string) (response, *apiError) { + ctx, cancel := a.dbContext(r.Context()) + defer cancel() + c, _, err := scanChat(a.db.QueryRow(ctx, chatSelect()+` WHERE c.id=$1 AND EXISTS(SELECT 1 FROM chat_members cm WHERE cm.chat_id=c.id AND cm.user_id=$2)`, chatID, uid)) + if errors.Is(err, pgx.ErrNoRows) { + return response{}, missing() + } + if err != nil { + return response{}, unavailable() + } + return response{status: 200, body: c}, nil +} + +func (a *API) listChats(r *http.Request, uid string) (response, *apiError) { + limit, afterTime, afterID, e := parsePage(r, "chats:"+uid) + if e != nil { + return response{}, e + } + ctx, cancel := a.dbContext(r.Context()) + defer cancel() + if afterTime != nil { + var ok bool + err := a.db.QueryRow(ctx, `SELECT EXISTS(SELECT 1 FROM chats c JOIN chat_members cm ON cm.chat_id=c.id WHERE cm.user_id=$1 AND c.id=$2 AND c.created_at=$3)`, uid, afterID, *afterTime).Scan(&ok) + if err != nil { + return response{}, unavailable() + } + if !ok { + return response{}, bad("Неверный cursor") + } + } + var rows pgx.Rows + var err error + base := chatSelect() + ` JOIN chat_members mine ON mine.chat_id=c.id WHERE mine.user_id=$1` + if afterTime == nil { + rows, err = a.db.Query(ctx, base+` ORDER BY c.created_at,c.id LIMIT $2`, uid, limit+1) + } else { + rows, err = a.db.Query(ctx, base+` AND (c.created_at,c.id)>($2,$3) ORDER BY c.created_at,c.id LIMIT $4`, uid, *afterTime, afterID, limit+1) + } + if err != nil { + return response{}, unavailable() + } + defer rows.Close() + items := make([]Chat, 0, limit+1) + times := make([]time.Time, 0, limit+1) + for rows.Next() { + c, t, scanErr := scanChat(rows) + if scanErr != nil { + return response{}, unavailable() + } + items = append(items, c) + times = append(times, t) + } + if rows.Err() != nil { + return response{}, unavailable() + } + var next any + if len(items) > limit { + last := items[limit-1] + next = cursorFor("chats:"+uid, times[limit-1], last.ID) + items = items[:limit] + } + return response{status: 200, body: map[string]any{"items": items, "next_cursor": next}}, nil +} + +func (a *API) createMessage(r *http.Request, uid, chatID string) (response, *apiError) { + key := r.Header.Get("Idempotency-Key") + if !idempotencyKeyRE.MatchString(key) { + return response{}, bad("Требуется корректный Idempotency-Key") + } + var in struct { + Text *string `json:"text"` + AttachmentIDs []string `json:"attachment_ids"` + } + if e := decode(r, &in); e != nil { + return response{}, e + } + if in.Text == nil { + return response{}, bad("Требуется поле text") + } + text := *in.Text + if !utf8.ValidString(text) || utf8.RuneCountInString(text) > 4000 || len(in.AttachmentIDs) > 10 || (strings.TrimSpace(text) == "" && len(in.AttachmentIDs) == 0) { + return response{}, bad("Текст сообщения должен содержать от 1 до 4000 символов") + } + if in.AttachmentIDs == nil { + in.AttachmentIDs = []string{} + } + seen := map[string]bool{} + for i, id := range in.AttachmentIDs { + id = strings.ToLower(id) + if !uuidRE.MatchString(id) || seen[id] { + return response{}, bad("Неверные attachment_ids") + } + seen[id] = true + in.AttachmentIDs[i] = id + } + ctx, cancel := a.dbContext(r.Context()) + defer cancel() + tx, err := a.db.Begin(ctx) + if err != nil { + return response{}, unavailable() + } + defer tx.Rollback(ctx) + var member bool + if err = tx.QueryRow(ctx, `SELECT EXISTS(SELECT 1 FROM chat_members WHERE chat_id=$1 AND user_id=$2)`, chatID, uid).Scan(&member); err != nil { + return response{}, unavailable() + } + if !member { + return response{}, missing() + } + requestHash := sha256.Sum256([]byte(text)) + if len(in.AttachmentIDs) > 0 { + canonical, _ := json.Marshal(in) + requestHash = sha256.Sum256(canonical) + } + if _, err = tx.Exec(ctx, `DELETE FROM message_idempotency WHERE user_id=$1 AND method=$2 AND path=$3 AND idempotency_key=$4 AND created_at < clock_timestamp() - ($5 * interval '1 second')`, uid, r.Method, r.URL.Path, key, int64(a.idempotencyTTL/time.Second)); err != nil { + return response{}, unavailable() + } + result, err := tx.Exec(ctx, `INSERT INTO message_idempotency(user_id,method,path,idempotency_key,request_hash) VALUES($1,$2,$3,$4,$5) ON CONFLICT DO NOTHING`, uid, r.Method, r.URL.Path, key, requestHash[:]) + if err != nil { + return response{}, unavailable() + } + inserted := result.RowsAffected() == 1 + if !inserted { + var storedHash []byte + var messageID string + if err = tx.QueryRow(ctx, `SELECT request_hash,message_id::text FROM message_idempotency WHERE user_id=$1 AND method=$2 AND path=$3 AND idempotency_key=$4 FOR UPDATE`, uid, r.Method, r.URL.Path, key).Scan(&storedHash, &messageID); err != nil { + return response{}, unavailable() + } + if !bytes.Equal(storedHash, requestHash[:]) { + return response{}, &apiError{status: 409, code: "conflict", message: "Idempotency-Key уже использован с другим запросом"} + } + var m Message + var created time.Time + if err = tx.QueryRow(ctx, `SELECT id::text,chat_id::text,sender_id::text,text,created_at,attachment_ids::text[] FROM messages WHERE id=$1`, messageID).Scan(&m.ID, &m.ChatID, &m.SenderID, &m.Text, &created, &m.AttachmentIDs); err != nil { + return response{}, unavailable() + } + m.CreatedAt = formatTime(created) + if err = tx.Commit(ctx); err != nil { + return response{}, unavailable() + } + return response{status: 201, body: m}, nil + } + for _, id := range in.AttachmentIDs { + var state string + err = tx.QueryRow(ctx, `SELECT status FROM attachments WHERE id=$1 AND chat_id=$2`, id, chatID).Scan(&state) + if errors.Is(err, pgx.ErrNoRows) { + return response{}, missing() + } + if err != nil { + return response{}, unavailable() + } + if state == "failed" { + return response{}, &apiError{status: 409, code: "conflict", message: "Вложение не готово"} + } + } + m := Message{ID: newID(), ChatID: chatID, SenderID: uid, Text: text, AttachmentIDs: in.AttachmentIDs} + var created time.Time + if err = tx.QueryRow(ctx, `INSERT INTO messages(id,chat_id,sender_id,text,attachment_ids) VALUES($1,$2,$3,$4,$5) RETURNING created_at`, m.ID, m.ChatID, m.SenderID, m.Text, m.AttachmentIDs).Scan(&created); err != nil { + return response{}, unavailable() + } + if _, err = tx.Exec(ctx, `UPDATE message_idempotency SET message_id=$1 WHERE user_id=$2 AND method=$3 AND path=$4 AND idempotency_key=$5`, m.ID, uid, r.Method, r.URL.Path, key); err != nil { + return response{}, unavailable() + } + if err = tx.Commit(ctx); err != nil { + return response{}, unavailable() + } + m.CreatedAt = formatTime(created) + return response{status: 201, body: m}, nil +} + +func (a *API) listMessages(r *http.Request, uid, chatID string) (response, *apiError) { + limit, afterTime, afterID, e := parsePage(r, "messages:"+chatID) + if e != nil { + return response{}, e + } + ctx, cancel := a.dbContext(r.Context()) + defer cancel() + var member bool + if err := a.db.QueryRow(ctx, `SELECT EXISTS(SELECT 1 FROM chat_members WHERE chat_id=$1 AND user_id=$2)`, chatID, uid).Scan(&member); err != nil { + return response{}, unavailable() + } + if !member { + return response{}, missing() + } + if afterTime != nil { + var ok bool + err := a.db.QueryRow(ctx, `SELECT EXISTS(SELECT 1 FROM messages WHERE chat_id=$1 AND id=$2 AND created_at=$3)`, chatID, afterID, *afterTime).Scan(&ok) + if err != nil { + return response{}, unavailable() + } + if !ok { + return response{}, bad("Неверный cursor") + } + } + var rows pgx.Rows + var err error + base := `SELECT id::text,chat_id::text,sender_id::text,text,created_at,attachment_ids::text[] FROM messages WHERE chat_id=$1` + if afterTime == nil { + rows, err = a.db.Query(ctx, base+` ORDER BY created_at,id LIMIT $2`, chatID, limit+1) + } else { + rows, err = a.db.Query(ctx, base+` AND (created_at,id)>($2,$3) ORDER BY created_at,id LIMIT $4`, chatID, *afterTime, afterID, limit+1) + } + if err != nil { + return response{}, unavailable() + } + defer rows.Close() + items := make([]Message, 0, limit+1) + times := make([]time.Time, 0, limit+1) + for rows.Next() { + var m Message + var t time.Time + if rows.Scan(&m.ID, &m.ChatID, &m.SenderID, &m.Text, &t, &m.AttachmentIDs) != nil { + return response{}, unavailable() + } + m.CreatedAt = formatTime(t) + items = append(items, m) + times = append(times, t) + } + if rows.Err() != nil { + return response{}, unavailable() + } + var next any + if len(items) > limit { + last := items[limit-1] + next = cursorFor("messages:"+chatID, times[limit-1], last.ID) + items = items[:limit] + } + return response{status: 200, body: map[string]any{"items": items, "next_cursor": next}}, nil +} diff --git a/internal/objectstore/bootstrap.go b/internal/objectstore/bootstrap.go new file mode 100644 index 0000000..4cb5e85 --- /dev/null +++ b/internal/objectstore/bootstrap.go @@ -0,0 +1,40 @@ +package objectstore + +import ( + "context" + "encoding/json" + "net/url" + + madmin "github.com/minio/madmin-go/v3" +) + +// Bootstrap receives root credentials only in the one-shot s3-init process. +// The API and cleaner receive a separate user limited to this bucket/prefix. +func Bootstrap(ctx context.Context, endpoint, bucket, rootAccess, rootSecret, apiAccess, apiSecret string) error { + u, err := url.Parse(endpoint) + if err != nil { + return err + } + admin, err := madmin.New(u.Host, rootAccess, rootSecret, u.Scheme == "https") + if err != nil { + return err + } + if err = admin.AddUser(ctx, apiAccess, apiSecret); err != nil { + return err + } + policy := map[string]any{ + "Version": "2012-10-17", + "Statement": []any{ + map[string]any{"Effect": "Allow", "Action": []string{"s3:GetBucketLocation", "s3:ListBucket"}, "Resource": []string{"arn:aws:s3:::" + bucket}}, + map[string]any{"Effect": "Allow", "Action": []string{"s3:GetObject", "s3:PutObject", "s3:DeleteObject"}, "Resource": []string{"arn:aws:s3:::" + bucket + "/originals/*", "arn:aws:s3:::" + bucket + "/staging/*", "arn:aws:s3:::" + bucket + "/variants/*"}}, + }, + } + encoded, err := json.Marshal(policy) + if err != nil { + return err + } + if err = admin.AddCannedPolicy(ctx, "messenger-api", encoded); err != nil { + return err + } + return admin.SetPolicy(ctx, "messenger-api", apiAccess, false) +} diff --git a/internal/objectstore/store.go b/internal/objectstore/store.go new file mode 100644 index 0000000..6866be2 --- /dev/null +++ b/internal/objectstore/store.go @@ -0,0 +1,108 @@ +package objectstore + +import ( + "context" + "fmt" + "io" + "net" + "net/http" + "net/url" + "time" + + "github.com/minio/minio-go/v7" + "github.com/minio/minio-go/v7/pkg/credentials" + "go.opentelemetry.io/otel/attribute" + "go.opentelemetry.io/otel/trace" + "mtusi-messenger/internal/telemetry" +) + +type Store struct { + Client *minio.Client + Bucket string + endpoint string +} + +func New(endpoint, region, bucket, access, secret string) (*Store, error) { + u, err := url.Parse(endpoint) + if err != nil || u.Host == "" || (u.Scheme != "http" && u.Scheme != "https") || u.Path != "" { + return nil, fmt.Errorf("invalid S3 endpoint") + } + transport := &http.Transport{ + DialContext: (&net.Dialer{Timeout: 2 * time.Second, KeepAlive: 30 * time.Second}).DialContext, + MaxIdleConns: 16, MaxIdleConnsPerHost: 8, MaxConnsPerHost: 8, + ResponseHeaderTimeout: 10 * time.Second, IdleConnTimeout: 30 * time.Second, + TLSHandshakeTimeout: 3 * time.Second, + } + client, err := minio.New(u.Host, &minio.Options{ + Creds: credentials.NewStaticV4(access, secret, ""), Secure: u.Scheme == "https", + Region: region, Transport: transport, MaxRetries: 1, + }) + if err != nil { + return nil, err + } + return &Store{Client: client, Bucket: bucket, endpoint: endpoint}, nil +} + +func (s *Store) Ready(ctx context.Context) (err error) { + ctx, span := telemetry.Span(ctx, "s3.bucket.exists", trace.SpanKindClient, attribute.String("db.system.name", "s3")) + defer func() { telemetry.Result(span, err) }() + ok, err := s.Client.BucketExists(ctx, s.Bucket) + if err != nil { + return err + } + if !ok { + return fmt.Errorf("S3 bucket missing") + } + return nil +} + +func (s *Store) Put(ctx context.Context, key string, r io.Reader, size int64, kind string) (err error) { + ctx, span := telemetry.Span(ctx, "s3.put", trace.SpanKindClient, attribute.String("db.system.name", "s3")) + defer func() { telemetry.Result(span, err) }() + _, err = s.Client.PutObject(ctx, s.Bucket, key, r, size, minio.PutObjectOptions{ + ContentType: kind, DisableMultipart: true, + }) + return err +} + +func (s *Store) Get(ctx context.Context, key string) (object *minio.Object, err error) { + ctx, span := telemetry.Span(ctx, "s3.get", trace.SpanKindClient, attribute.String("db.system.name", "s3")) + defer func() { telemetry.Result(span, err) }() + object, err = s.Client.GetObject(ctx, s.Bucket, key, minio.GetObjectOptions{}) + if err != nil { + return nil, err + } + if _, err = object.Stat(); err != nil { + object.Close() + return nil, err + } + return object, nil +} + +func (s *Store) Delete(ctx context.Context, key string) (err error) { + ctx, span := telemetry.Span(ctx, "s3.delete", trace.SpanKindClient, attribute.String("db.system.name", "s3")) + defer func() { telemetry.Result(span, err) }() + return s.Client.RemoveObject(ctx, s.Bucket, key, minio.RemoveObjectOptions{}) +} + +// Audit performs an actual unauthenticated GET and an out-of-scope operation +// with API credentials. Neither check prints credentials or object URLs. +func (s *Store) Audit(ctx context.Context, key string) error { + req, err := http.NewRequestWithContext(ctx, "GET", s.endpoint+"/"+s.Bucket+"/"+key, nil) + if err != nil { + return err + } + response, err := (&http.Client{Timeout: 5 * time.Second}).Do(req) + if err != nil { + return err + } + response.Body.Close() + if response.StatusCode != 403 { + return fmt.Errorf("anonymous GET returned %d", response.StatusCode) + } + err = s.Client.MakeBucket(ctx, "forbidden-audit-bucket", minio.MakeBucketOptions{}) + if minio.ToErrorResponse(err).Code != "AccessDenied" { + return fmt.Errorf("API credentials are not restricted: %v", err) + } + return nil +} diff --git a/internal/pipeline/pipeline.go b/internal/pipeline/pipeline.go new file mode 100644 index 0000000..e41a2b2 --- /dev/null +++ b/internal/pipeline/pipeline.go @@ -0,0 +1,549 @@ +// Package pipeline implements durable intent, at-least-once delivery and fenced effects. +package pipeline + +import ( + "context" + "crypto/sha256" + "encoding/hex" + "encoding/json" + "errors" + "fmt" + "io" + "log" + "math/rand/v2" + "os" + "os/exec" + "path/filepath" + "strconv" + "time" + + "github.com/google/uuid" + "github.com/jackc/pgx/v5" + "github.com/jackc/pgx/v5/pgxpool" + "github.com/redis/go-redis/v9" + "go.opentelemetry.io/otel/attribute" + "go.opentelemetry.io/otel/codes" + "go.opentelemetry.io/otel/trace" + "mtusi-messenger/internal/objectstore" + "mtusi-messenger/internal/telemetry" +) + +const Stream = "attachment.jobs.v1" +const Group = "image-workers" + +type Pipeline struct { + DB *pgxpool.Pool + Broker *redis.Client + Objects *objectstore.Store + Timeout time.Duration + Retention time.Duration + Logger *log.Logger +} +type Job struct { + SchemaVersion int `json:"schema_version"` + ID string `json:"job_id"` + Type string `json:"type"` + AttachmentID string `json:"attachment_id"` + Version int `json:"pipeline_version"` + Attempt int `json:"attempt"` + Created string `json:"created_at"` + Traceparent string `json:"traceparent,omitempty"` + Tracestate string `json:"tracestate,omitempty"` +} +type Variant struct { + Name string `json:"name"` + ContentType string `json:"content_type"` + Size int64 `json:"size_bytes"` + Width int `json:"width"` + Height int `json:"height"` + SHA string `json:"sha256"` +} + +func (p *Pipeline) Checkpoint(ctx context.Context, point string) { + var armed bool + err := p.DB.QueryRow(ctx, `UPDATE pipeline_faults SET armed=false,hit_at=clock_timestamp() WHERE point=$1 AND armed RETURNING true`, point).Scan(&armed) + if err == nil && armed { + p.Logger.Printf("injected crash: %s", point) + os.Exit(86) + } +} +func pause(ctx context.Context, d time.Duration) { + select { + case <-ctx.Done(): + case <-time.After(d): + } +} + +// Dispatch re-publishes unfinished intents if a broker confirmation/connection is lost. +// Publication is deliberately outside the PostgreSQL transaction; duplicates are safe. +func (p *Pipeline) Dispatch(ctx context.Context) { + for ctx.Err() == nil { + if err := p.dispatchOnce(ctx); err != nil { + p.Logger.Print("dispatcher unavailable; intent retained") + } + sweepctx, cancelSweep := context.WithTimeout(ctx, 5*time.Second) + if err := p.Sweep(sweepctx); err != nil { + p.Logger.Print("staging sweep deferred") + } + cancelSweep() + pause(ctx, time.Second) + } +} +func (p *Pipeline) dispatchOnce(ctx context.Context) error { + rows, err := p.DB.Query(ctx, `SELECT j.id::text,j.attachment_id::text,j.pipeline_version,j.attempts+1,j.created_at,j.traceparent,j.tracestate + FROM attachment_jobs j JOIN attachment_outbox o ON o.job_id=j.id + WHERE ((j.state='queued' AND j.next_at<=clock_timestamp()) OR (j.state='processing' AND j.lease_until 4 { + return fmt.Errorf("WORKER_CONCURRENCY must be 1..4") + } + // One message at a time per consumer: prefetch=1. + done := make(chan struct{}, concurrency) + for i := 0; i < concurrency; i++ { + go func() { defer func() { done <- struct{}{} }(); p.consume(ctx, uuid.NewString()) }() + } + for i := 0; i < concurrency; i++ { + <-done + } + return nil +} +func (p *Pipeline) consume(ctx context.Context, consumer string) { + reclaimIdle := p.Timeout + 10*time.Second + for ctx.Err() == nil { + _ = p.Broker.XGroupCreateMkStream(ctx, Stream, Group, "0").Err() + pending, _, err := p.Broker.XAutoClaim(ctx, &redis.XAutoClaimArgs{Stream: Stream, Group: Group, Consumer: consumer, MinIdle: reclaimIdle, Start: "0-0", Count: 1}).Result() + if err == nil && len(pending) > 0 { + p.handle(ctx, pending[0]) + continue + } + messages, err := p.Broker.XReadGroup(ctx, &redis.XReadGroupArgs{Group: Group, Consumer: consumer, Streams: []string{Stream, ">"}, Count: 1, Block: time.Second}).Result() + if err != nil { + pause(ctx, 200*time.Millisecond) + continue + } + for _, stream := range messages { + for _, m := range stream.Messages { + p.handle(ctx, m) + } + } + } +} +func (p *Pipeline) handle(ctx context.Context, m redis.XMessage) { + raw, ok := m.Values["job"].(string) + var j Job + if !ok || json.Unmarshal([]byte(raw), &j) != nil || j.SchemaVersion != 1 || j.Version != 1 || j.Type != "attachment.process.v1" || uuid.Validate(j.ID) != nil || uuid.Validate(j.AttachmentID) != nil { + // Malformed envelopes go to a diagnostic dead-letter stream, never interpreted as storage paths. + if p.Broker.XAdd(ctx, &redis.XAddArgs{Stream: "attachment.invalid.dlq", Values: map[string]any{"stream_id": m.ID, "error_code": "invalid_envelope"}}).Err() == nil { + p.ack(ctx, m.ID) + } + return + } + + ctx = telemetry.Extract(ctx, j.Traceparent, j.Tracestate) + ctx, consumer := telemetry.Span(ctx, "attachment.consume", trace.SpanKindConsumer, attribute.String("messaging.system", "redis"), attribute.String("messaging.destination.name", Stream), attribute.String("job.id", j.ID)) + began := time.Now() + attempted := false + outcome := "abandoned" + defer func() { + consumer.SetAttributes(attribute.String("job.outcome", outcome)) + if outcome == "retry" || outcome == "failed" || outcome == "abandoned" { + consumer.SetStatus(codes.Error, outcome) + } + if attempted { + telemetry.Busy.Add(-1) + telemetry.Job(ctx, outcome, time.Since(began)) + } + telemetry.Event(ctx, "job.completed", map[string]any{"job_id": j.ID, "outcome": outcome, "duration_ms": time.Since(began).Milliseconds()}) + consumer.End() + }() + jobctx, cancel := context.WithTimeout(ctx, p.Timeout) + defer cancel() + conn, err := p.DB.Acquire(jobctx) + if err != nil { + return + } + defer conn.Release() + var locked bool + if err = conn.QueryRow(jobctx, `SELECT pg_try_advisory_lock(hashtextextended($1,5))`, j.ID).Scan(&locked); err != nil || !locked { + return + } + defer func() { + unlockctx, done := context.WithTimeout(context.Background(), 3*time.Second) + defer done() + _, e := conn.Exec(unlockctx, `SELECT pg_advisory_unlock(hashtextextended($1,5))`, j.ID) + if e != nil { + _ = conn.Conn().Close(unlockctx) + } + }() + tx, err := conn.Begin(jobctx) + if err != nil { + return + } + defer tx.Rollback(context.Background()) + var state, staging, attachment, requestID string + var accepted time.Time + var attempt, maxAttempts int + var due bool + err = tx.QueryRow(jobctx, `SELECT state,staging_key,attachment_id::text,attempts,max_attempts, + (state='queued' AND next_at<=clock_timestamp()) OR (state='processing' AND lease_until= maxAttempts { + if p.failTx(jobctx, tx, j.ID, attachment, "retry_exhausted") == nil && tx.Commit(jobctx) == nil { + p.ack(ctx, m.ID) + } + return + } + token := uuid.NewString() + consumer.SetAttributes(attribute.Int("job.attempt", attempt+1)) + attempt++ + _, err = tx.Exec(jobctx, `UPDATE attachment_jobs SET state='processing',attempts=$2,lease_token=$3,lease_until=clock_timestamp()+($4*interval '1 second'),updated_at=clock_timestamp() WHERE id=$1`, j.ID, attempt, token, int(p.Timeout/time.Second)+5) + if err != nil { + return + } + if _, err = tx.Exec(jobctx, `UPDATE attachments SET status='processing',updated_at=clock_timestamp() WHERE id=$1`, attachment); err != nil { + return + } + if tx.Commit(jobctx) != nil { + return + } + attempted = true + telemetry.Busy.Add(1) + telemetry.Event(jobctx, "job.started", map[string]any{"job_id": j.ID, "attempt": attempt}) + p.Checkpoint(jobctx, "worker_after_claim") + variants, code := p.process(jobctx, attachment, staging) + p.Checkpoint(jobctx, "worker_after_objects") + // A fresh transaction completes after job timeout/cancel; token fences stale attempts. + finishctx, finishcancel := context.WithTimeout(context.WithoutCancel(ctx), 5*time.Second) + defer finishcancel() + finish, err := conn.Begin(finishctx) + if err != nil { + return + } + defer finish.Rollback(context.Background()) + var current string + if err = finish.QueryRow(finishctx, `SELECT lease_token::text FROM attachment_jobs WHERE id=$1 AND state='processing' FOR UPDATE`, j.ID).Scan(¤t); err != nil || current != token { + return + } + if code == "" { + outcome = "success" + encoded, _ := json.Marshal(variants) + _, err = finish.Exec(finishctx, `UPDATE attachments SET status='ready',object_key=$2,variants=$3,error_code=NULL,updated_at=clock_timestamp() WHERE id=$1`, attachment, "originals/"+attachment, encoded) + if err == nil { + _, err = finish.Exec(finishctx, `UPDATE attachment_jobs SET state='ready',lease_until=NULL,lease_token=NULL,error_code=NULL,updated_at=clock_timestamp() WHERE id=$1`, j.ID) + } + } else if code == "invalid_image" || attempt >= maxAttempts { + outcome = "failed" + if code != "invalid_image" { + code = "retry_exhausted" + } + err = p.failTx(finishctx, finish, j.ID, attachment, code) + } else { + outcome = "retry" + delay := RetryDelay(attempt) + _, err = finish.Exec(finishctx, `UPDATE attachment_jobs SET state='queued',error_code=$2,next_at=clock_timestamp()+($3*interval '1 millisecond'),lease_until=NULL,lease_token=NULL,updated_at=clock_timestamp() WHERE id=$1`, j.ID, code, delay.Milliseconds()) + if err == nil { + _, err = finish.Exec(finishctx, `UPDATE attachments SET status='queued',updated_at=clock_timestamp() WHERE id=$1`, attachment) + } + } + if err != nil || finish.Commit(finishctx) != nil { + outcome = "abandoned" + return + } + if outcome == "success" { + telemetry.Ready(ctx, time.Since(accepted)) + } + p.Checkpoint(finishctx, "worker_before_ack") + p.ack(finishctx, m.ID) +} +func RetryDelay(attempt int) time.Duration { + if attempt > 6 { + attempt = 6 + } + base := time.Second * time.Duration(1< 10485760 { + return nil, "invalid_image" + } + var expected string + if p.DB.QueryRow(ctx, `SELECT sha256 FROM attachments WHERE id=$1`, id).Scan(&expected) != nil { + return nil, "storage_unavailable" + } + if hex.EncodeToString(hash.Sum(nil)) != expected { + return nil, "invalid_image" + } + cmd := exec.CommandContext(ctx, "python3", "/opt/process_image.py", filepath.Join(dir, "source"), dir) + _, imagespan := telemetry.Span(ctx, "image.process", trace.SpanKindInternal) + output, err := cmd.Output() + telemetry.Result(imagespan, err) + if err != nil { + if ctx.Err() != nil { + return nil, "processing_timeout" + } + var exit *exec.ExitError + if errors.As(err, &exit) && exit.ExitCode() == 2 { + return nil, "invalid_image" + } + return nil, "processing_failed" + } + variants := []Variant{} + if json.Unmarshal(output, &variants) != nil || len(variants) != 3 { + return nil, "processing_failed" + } + var kind string + if p.DB.QueryRow(ctx, `SELECT content_type FROM attachments WHERE id=$1`, id).Scan(&kind) != nil { + return nil, "storage_unavailable" + } + source, err := os.Open(filepath.Join(dir, "source")) + if err != nil { + return nil, "storage_unavailable" + } + err = p.Objects.Put(ctx, "originals/"+id, source, n, kind) + source.Close() + if err != nil { + return nil, "storage_unavailable" + } + for _, v := range variants { + f, err := os.Open(filepath.Join(dir, v.Name)) + if err != nil { + return nil, "processing_failed" + } + err = p.Objects.Put(ctx, "variants/"+id+"/v1/"+v.Name, f, v.Size, v.ContentType) + f.Close() + if err != nil { + return nil, "storage_unavailable" + } + } + return variants, "" +} + +// Failed registry is PostgreSQL; explicit replay is audited and refuses successful jobs. +func (p *Pipeline) Replay(ctx context.Context, id string) error { + if uuid.Validate(id) != nil { + return fmt.Errorf("invalid job id") + } + conn, err := p.DB.Acquire(ctx) + if err != nil { + return err + } + defer conn.Release() + tx, err := conn.Begin(ctx) + if err != nil { + return err + } + defer tx.Rollback(ctx) + var attachment, state, code string + var retained bool + err = tx.QueryRow(ctx, `SELECT attachment_id::text,state,COALESCE(error_code,''),staging_deleted_at IS NULL FROM attachment_jobs WHERE id=$1 FOR UPDATE`, id).Scan(&attachment, &state, &code, &retained) + if err != nil { + return err + } + if state != "failed" || !retained { + return fmt.Errorf("only failed jobs with retained staging can be replayed") + } + var key string + if err = tx.QueryRow(ctx, `SELECT staging_key FROM attachment_jobs WHERE id=$1`, id).Scan(&key); err != nil { + return err + } + object, err := p.Objects.Get(ctx, key) + if err != nil { + return fmt.Errorf("staging unavailable") + } + object.Close() + if _, err = tx.Exec(ctx, `INSERT INTO attachment_replays(job_id,previous_error) VALUES($1,$2)`, id, code); err != nil { + return err + } + if _, err = tx.Exec(ctx, `UPDATE attachment_jobs SET state='queued',attempts=0,error_code=NULL,next_at=clock_timestamp(),lease_until=NULL,lease_token=NULL,updated_at=clock_timestamp() WHERE id=$1`, id); err != nil { + return err + } + if _, err = tx.Exec(ctx, `UPDATE attachments SET status='queued',error_code=NULL,updated_at=clock_timestamp() WHERE id=$1`, attachment); err != nil { + return err + } + if _, err = tx.Exec(ctx, `UPDATE attachment_outbox SET published_at=NULL WHERE job_id=$1`, id); err != nil { + return err + } + return tx.Commit(ctx) +} + +func (p *Pipeline) Sweep(ctx context.Context) error { + rows, err := p.DB.Query(ctx, `SELECT id::text,staging_key FROM attachment_jobs WHERE staging_deleted_at IS NULL AND (state='ready' OR (state='failed' AND updated_at 6 { + bounded = 6 + } + base := time.Second * time.Duration(1< base+base/2 { + t.Fatalf("attempt %d: delay %s outside [%s,%s]", attempt, got, base, base+base/2) + } + } + } +} diff --git a/internal/telemetry/dependencies.go b/internal/telemetry/dependencies.go new file mode 100644 index 0000000..8d9d110 --- /dev/null +++ b/internal/telemetry/dependencies.go @@ -0,0 +1,74 @@ +package telemetry + +import ( + "context" + "github.com/jackc/pgx/v5" + "github.com/jackc/pgx/v5/pgxpool" + "github.com/redis/go-redis/v9" + "go.opentelemetry.io/otel/attribute" + "go.opentelemetry.io/otel/trace" + "net" + "strings" +) + +// Queries are named by operation only. SQL text/args never enter telemetry. +type DBTracer struct{} + +func (DBTracer) TraceQueryStart(ctx context.Context, _ *pgx.Conn, d pgx.TraceQueryStartData) context.Context { + if !trace.SpanContextFromContext(ctx).IsValid() { + return ctx + } + op := "OTHER" + f := strings.Fields(d.SQL) + if len(f) > 0 { + switch strings.ToUpper(f[0]) { + case "SELECT", "INSERT", "UPDATE", "DELETE", "BEGIN", "COMMIT", "ROLLBACK", "CREATE", "ALTER": + op = strings.ToUpper(f[0]) + } + } + c, _ := Span(ctx, "db."+strings.ToLower(op), trace.SpanKindClient, attribute.String("db.system.name", "postgresql"), attribute.String("db.operation.name", op)) + return c +} +func (DBTracer) TraceQueryEnd(ctx context.Context, _ *pgx.Conn, d pgx.TraceQueryEndData) { + if trace.SpanContextFromContext(ctx).IsValid() { + Result(trace.SpanFromContext(ctx), d.Err) + } +} +func (DBTracer) TraceAcquireStart(ctx context.Context, _ *pgxpool.Pool, _ pgxpool.TraceAcquireStartData) context.Context { + Waiting.Add(1) + if trace.SpanContextFromContext(ctx).IsValid() { + c, _ := Span(ctx, "db.pool.acquire", trace.SpanKindClient, attribute.String("db.system.name", "postgresql")) + return c + } + return ctx +} +func (DBTracer) TraceAcquireEnd(ctx context.Context, _ *pgxpool.Pool, data pgxpool.TraceAcquireEndData) { + Waiting.Add(-1) + if trace.SpanContextFromContext(ctx).IsValid() { + Result(trace.SpanFromContext(ctx), data.Err) + } +} + +type RedisHook struct{} + +func (RedisHook) DialHook(next redis.DialHook) redis.DialHook { + return func(ctx context.Context, network, addr string) (net.Conn, error) { return next(ctx, network, addr) } +} +func (RedisHook) ProcessHook(next redis.ProcessHook) redis.ProcessHook { + return func(ctx context.Context, cmd redis.Cmder) error { + if !trace.SpanContextFromContext(ctx).IsValid() { + return next(ctx, cmd) + } + c, s := Span(ctx, "redis."+cmd.Name(), trace.SpanKindClient, attribute.String("db.system.name", "redis"), attribute.String("db.operation.name", cmd.Name())) + err := next(c, cmd) + if err == redis.Nil { + s.End() + } else { + Result(s, err) + } + return err + } +} +func (RedisHook) ProcessPipelineHook(next redis.ProcessPipelineHook) redis.ProcessPipelineHook { + return next +} diff --git a/internal/telemetry/metrics.go b/internal/telemetry/metrics.go new file mode 100644 index 0000000..feac579 --- /dev/null +++ b/internal/telemetry/metrics.go @@ -0,0 +1,93 @@ +package telemetry + +import ( + "context" + "github.com/jackc/pgx/v5/pgxpool" + "go.opentelemetry.io/otel" + "go.opentelemetry.io/otel/attribute" + "go.opentelemetry.io/otel/metric" + "os" + "runtime" + "strconv" + "strings" + "syscall" + "time" +) + +func Monitor(pool *pgxpool.Pool, role string, capacity int) error { + m := otel.Meter(Scope) + names := []string{"messenger_db_pool_used", "messenger_db_pool_idle", "messenger_db_pool_waiting", "messenger_db_pool_capacity", "messenger_worker_busy", "messenger_worker_capacity", "messenger_queue_depth", "messenger_queue_oldest_age_seconds", "messenger_db_available", "messenger_process_memory_bytes", "messenger_container_memory_bytes", "messenger_container_memory_limit_bytes", "messenger_telemetry_export_errors", "messenger_process_last_observed_seconds"} + gauges := map[string]metric.Int64ObservableGauge{} + for _, n := range names { + g, err := m.Int64ObservableGauge(n) + if err != nil { + return err + } + gauges[n] = g + } + containerCPU, _ := m.Float64ObservableCounter("messenger_container_cpu", metric.WithUnit("s")) + cpu, _ := m.Float64ObservableCounter("messenger_process_cpu", metric.WithUnit("s")) + acquired, _ := m.Int64ObservableCounter("messenger_db_pool_acquire_wait") + duration, _ := m.Float64ObservableCounter("messenger_db_pool_acquire_wait_duration", metric.WithUnit("s")) + instruments := []metric.Observable{cpu, containerCPU, acquired, duration} + for _, g := range gauges { + instruments = append(instruments, g) + } + _, err := m.RegisterCallback(func(ctx context.Context, o metric.Observer) error { + stat := pool.Stat() + observeTime := time.Now().Unix() + observe := func(n string, value int64) { o.ObserveInt64(gauges[n], value) } + observe("messenger_db_pool_used", int64(stat.AcquiredConns())) + observe("messenger_db_pool_idle", int64(stat.IdleConns())) + observe("messenger_db_pool_waiting", Waiting.Load()) + observe("messenger_db_pool_capacity", int64(stat.MaxConns())) + o.ObserveInt64(acquired, stat.EmptyAcquireCount()) + o.ObserveFloat64(duration, stat.EmptyAcquireWaitTime().Seconds()) + observe("messenger_telemetry_export_errors", ExportErrors.Load()) + observe("messenger_process_last_observed_seconds", observeTime) + if role == "worker" { + observe("messenger_worker_busy", Busy.Load()) + observe("messenger_worker_capacity", int64(capacity)) + } + var mem runtime.MemStats + runtime.ReadMemStats(&mem) + observe("messenger_process_memory_bytes", int64(mem.Sys)) + for file, name := range map[string]string{"memory.current": "messenger_container_memory_bytes", "memory.max": "messenger_container_memory_limit_bytes"} { + if b, e := os.ReadFile("/sys/fs/cgroup/" + file); e == nil { + if n, e := strconv.ParseInt(strings.TrimSpace(string(b)), 10, 64); e == nil { + observe(name, n) + } + } + } + + if raw, e := os.ReadFile("/sys/fs/cgroup/cpu.stat"); e == nil { + for _, line := range strings.Split(string(raw), "\n") { + fields := strings.Fields(line) + if len(fields) == 2 && fields[0] == "usage_usec" { + if n, e := strconv.ParseFloat(fields[1], 64); e == nil { + o.ObserveFloat64(containerCPU, n/1e6) + } + } + } + } + var usage syscall.Rusage + if syscall.Getrusage(syscall.RUSAGE_SELF, &usage) == nil { + o.ObserveFloat64(cpu, float64(usage.Utime.Sec+usage.Stime.Sec)+float64(usage.Utime.Usec+usage.Stime.Usec)/1e6) + } + if role == "dispatcher" { + c, cancel := context.WithTimeout(ctx, time.Second) + defer cancel() + var depth, age int64 + e := pool.QueryRow(c, `SELECT count(*),COALESCE(EXTRACT(EPOCH FROM clock_timestamp()-min(created_at))::bigint,0) FROM attachment_jobs WHERE state='queued'`).Scan(&depth, &age) + if e == nil { + o.ObserveInt64(gauges["messenger_queue_depth"], depth, metric.WithAttributes(attribute.String("queue", "attachment.jobs.v1"))) + o.ObserveInt64(gauges["messenger_queue_oldest_age_seconds"], age, metric.WithAttributes(attribute.String("queue", "attachment.jobs.v1"))) + observe("messenger_db_available", 1) + } else { + observe("messenger_db_available", 0) + } + } + return nil + }, instruments...) + return err +} diff --git a/internal/telemetry/telemetry.go b/internal/telemetry/telemetry.go new file mode 100644 index 0000000..8229637 --- /dev/null +++ b/internal/telemetry/telemetry.go @@ -0,0 +1,205 @@ +// Package telemetry exports bounded, asynchronous OTLP signals without business payloads. +package telemetry + +import ( + "context" + "encoding/json" + "fmt" + "os" + "strings" + "sync/atomic" + "time" + + "go.opentelemetry.io/otel" + "go.opentelemetry.io/otel/attribute" + "go.opentelemetry.io/otel/codes" + "go.opentelemetry.io/otel/exporters/otlp/otlplog/otlploghttp" + "go.opentelemetry.io/otel/exporters/otlp/otlpmetric/otlpmetrichttp" + "go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp" + otellog "go.opentelemetry.io/otel/log" + logglobal "go.opentelemetry.io/otel/log/global" + "go.opentelemetry.io/otel/metric" + "go.opentelemetry.io/otel/propagation" + sdklog "go.opentelemetry.io/otel/sdk/log" + sdkmetric "go.opentelemetry.io/otel/sdk/metric" + "go.opentelemetry.io/otel/sdk/resource" + sdktrace "go.opentelemetry.io/otel/sdk/trace" + "go.opentelemetry.io/otel/trace" +) + +const Scope = "mtusi-messenger/telemetry-v1" + +var Service = "messenger-api" +var Busy atomic.Int64 +var Waiting atomic.Int64 +var ExportErrors atomic.Int64 +var signals *Signals + +type Signals struct { + Requests metric.Int64Counter + HTTPDuration metric.Float64Histogram + Jobs metric.Int64Counter + JobDuration metric.Float64Histogram + ReadyDuration metric.Float64Histogram +} + +func Start(ctx context.Context, service, instance string) (func(context.Context) error, error) { + Service = service + otel.SetTextMapPropagator(propagation.TraceContext{}) + attrs := []attribute.KeyValue{attribute.String("service.name", service), attribute.String("service.version", "lab6-v1"), attribute.String("service.instance.id", instance), attribute.String("deployment.environment.name", "lab")} + res := resource.NewWithAttributes("https://opentelemetry.io/schemas/1.37.0", attrs...) + traceExp, err := otlptracehttp.New(ctx, otlptracehttp.WithTimeout(2*time.Second), otlptracehttp.WithRetry(otlptracehttp.RetryConfig{Enabled: false})) + if err != nil { + return nil, err + } + metricExp, err := otlpmetrichttp.New(ctx, otlpmetrichttp.WithTimeout(2*time.Second), otlpmetrichttp.WithRetry(otlpmetrichttp.RetryConfig{Enabled: false})) + if err != nil { + return nil, err + } + logExp, err := otlploghttp.New(ctx, otlploghttp.WithTimeout(2*time.Second), otlploghttp.WithRetry(otlploghttp.RetryConfig{Enabled: false})) + if err != nil { + return nil, err + } + tp := sdktrace.NewTracerProvider(sdktrace.WithResource(res), sdktrace.WithSampler(sdktrace.ParentBased(sdktrace.AlwaysSample())), sdktrace.WithBatcher(traceExp, sdktrace.WithMaxQueueSize(256), sdktrace.WithMaxExportBatchSize(64), sdktrace.WithBatchTimeout(500*time.Millisecond), sdktrace.WithExportTimeout(2*time.Second))) + mp := sdkmetric.NewMeterProvider(sdkmetric.WithResource(res), sdkmetric.WithReader(sdkmetric.NewPeriodicReader(metricExp, sdkmetric.WithInterval(2*time.Second), sdkmetric.WithTimeout(2*time.Second)))) + lp := sdklog.NewLoggerProvider(sdklog.WithResource(res), sdklog.WithProcessor(sdklog.NewBatchProcessor(logExp, sdklog.WithMaxQueueSize(256), sdklog.WithExportMaxBatchSize(64), sdklog.WithExportInterval(500*time.Millisecond), sdklog.WithExportTimeout(2*time.Second)))) + otel.SetTracerProvider(tp) + otel.SetMeterProvider(mp) + logglobal.SetLoggerProvider(lp) + // No exporter/network details, payloads or credentials are logged on errors. + otel.SetErrorHandler(otel.ErrorHandlerFunc(func(error) { ExportErrors.Add(1) })) + meter := otel.Meter(Scope) + s := &Signals{} + s.Requests, err = meter.Int64Counter("messenger_http_requests") + if err != nil { + return nil, err + } + buckets := []float64{.005, .01, .025, .05, .1, .25, .5, 1, 2.5, 5, 10, 30, 60, 120} + s.HTTPDuration, _ = meter.Float64Histogram("messenger_http_duration", metric.WithUnit("s"), metric.WithExplicitBucketBoundaries(buckets...)) + s.Jobs, _ = meter.Int64Counter("messenger_jobs_processed") + s.JobDuration, _ = meter.Float64Histogram("messenger_job_duration", metric.WithUnit("s"), metric.WithExplicitBucketBoundaries(buckets...)) + s.ReadyDuration, _ = meter.Float64Histogram("messenger_time_to_ready", metric.WithUnit("s"), metric.WithExplicitBucketBoundaries(.1, .5, 1, 5, 10, 30, 60, 120, 300, 900, 3600, 86400)) + signals = s + return func(c context.Context) error { + // Shared deadline bounds shutdown across all signals. + a := tp.Shutdown(c) + b := mp.Shutdown(c) + d := lp.Shutdown(c) + if a != nil || b != nil || d != nil { + return fmt.Errorf("telemetry shutdown incomplete") + } + return nil + }, nil +} +func Span(ctx context.Context, name string, kind trace.SpanKind, attrs ...attribute.KeyValue) (context.Context, trace.Span) { + return otel.Tracer(Scope).Start(ctx, name, trace.WithSpanKind(kind), trace.WithAttributes(attrs...)) +} +func Result(span trace.Span, err error) { + if err != nil { + span.SetStatus(codes.Error, "dependency unavailable") + span.SetAttributes(attribute.Bool("error", true)) + } + span.End() +} +func Inject(ctx context.Context) (string, string) { + c := propagation.MapCarrier{} + otel.GetTextMapPropagator().Inject(ctx, c) + return c.Get("traceparent"), c.Get("tracestate") +} +func Extract(ctx context.Context, parent, state string) context.Context { + return otel.GetTextMapPropagator().Extract(ctx, propagation.MapCarrier{"traceparent": parent, "tracestate": state}) +} + +type requestKey struct{} + +func WithRequestID(ctx context.Context, id string) context.Context { + return context.WithValue(ctx, requestKey{}, id) +} +func RequestID(ctx context.Context) string { v, _ := ctx.Value(requestKey{}).(string); return v } +func Event(ctx context.Context, event string, fields map[string]any) { + sc := trace.SpanContextFromContext(ctx) + body := map[string]any{"timestamp": time.Now().UTC().Format(time.RFC3339Nano), "severity": "INFO", "service.name": Service, "event": event} + if sc.IsValid() { + body["trace_id"] = sc.TraceID().String() + body["span_id"] = sc.SpanID().String() + } + if id, ok := ctx.Value(requestKey{}).(string); ok { + body["request_id"] = id + } + for k, v := range fields { + body[k] = v + } + severity := otellog.SeverityInfo + if status, ok := fields["status"].(int); ok && status >= 500 { + severity = otellog.SeverityError + } + if outcome, ok := fields["outcome"].(string); ok && (outcome == "failed" || outcome == "retry" || outcome == "abandoned") { + severity = otellog.SeverityError + } + if severity == otellog.SeverityError { + body["severity"] = "ERROR" + } + encoded, _ := json.Marshal(body) + _, _ = os.Stdout.Write(append(encoded, '\n')) + var record otellog.Record + record.SetTimestamp(time.Now()) + record.SetSeverity(severity) + record.SetSeverityText(body["severity"].(string)) + record.SetBody(otellog.StringValue(string(encoded))) + record.AddAttributes(otellog.String("event", event)) + logglobal.GetLoggerProvider().Logger(Scope).Emit(ctx, record) +} +func HTTP(ctx context.Context, method, route string, status int, elapsed time.Duration) { + if signals == nil { + return + } + a := metric.WithAttributes(attribute.String("method", method), attribute.String("route", route), attribute.String("status_class", fmt.Sprintf("%dxx", status/100))) + signals.Requests.Add(ctx, 1, a) + signals.HTTPDuration.Record(ctx, elapsed.Seconds(), a) +} +func Job(ctx context.Context, outcome string, d time.Duration) { + if signals == nil { + return + } + a := metric.WithAttributes(attribute.String("job_type", "attachment.process.v1"), attribute.String("outcome", outcome)) + signals.Jobs.Add(ctx, 1, a) + signals.JobDuration.Record(ctx, d.Seconds(), a) +} +func Ready(ctx context.Context, d time.Duration) { + if signals != nil { + signals.ReadyDuration.Record(ctx, d.Seconds(), metric.WithAttributes(attribute.String("job_type", "attachment.process.v1"))) + } +} +func Method(s string) string { + switch s { + case "GET", "POST", "DELETE", "PUT", "PATCH", "HEAD", "OPTIONS": + return s + } + return "OTHER" +} +func Route(path string) string { + switch path { + case "/health/live", "/health/ready", "/api/v1/users", "/api/v1/users/me", "/api/v1/auth/sessions", "/api/v1/auth/sessions/current", "/api/v1/chats": + return path + } + parts := strings.Split(strings.Trim(path, "/"), "/") + if len(parts) >= 4 && parts[0] == "api" && parts[1] == "v1" { + if parts[2] == "chats" { + if len(parts) == 4 { + return "/api/v1/chats/{chat_id}" + } + if len(parts) == 5 && (parts[4] == "messages" || parts[4] == "attachments") { + return "/api/v1/chats/{chat_id}/" + parts[4] + } + } + if parts[2] == "attachments" { + if len(parts) == 4 { + return "/api/v1/attachments/{attachment_id}" + } + if len(parts) == 5 && parts[4] == "content" { + return "/api/v1/attachments/{attachment_id}/content" + } + } + } + return "/unmatched" +} diff --git a/internal/telemetry/telemetry_test.go b/internal/telemetry/telemetry_test.go new file mode 100644 index 0000000..0e934f9 --- /dev/null +++ b/internal/telemetry/telemetry_test.go @@ -0,0 +1,28 @@ +package telemetry + +import ( + "strings" + "testing" +) + +func TestRouteLabelsNeverContainResourceIDs(t *testing.T) { + ids := []string{"11111111-1111-4111-8111-111111111111", "22222222-2222-4222-8222-222222222222"} + for _, id := range ids { + cases := map[string]string{ + "/api/v1/chats/" + id: "/api/v1/chats/{chat_id}", + "/api/v1/chats/" + id + "/messages": "/api/v1/chats/{chat_id}/messages", + "/api/v1/chats/" + id + "/attachments": "/api/v1/chats/{chat_id}/attachments", + "/api/v1/attachments/" + id + "/content": "/api/v1/attachments/{attachment_id}/content", + "/arbitrary/" + id: "/unmatched", + } + for input, want := range cases { + got := Route(input) + if got != want || strings.Contains(got, id) { + t.Fatalf("unsafe route label %q for %q", got, input) + } + } + } + if Method("CUSTOM-USER-METHOD") != "OTHER" { + t.Fatal("unbounded method label") + } +} diff --git a/scripts/init-local.sh b/scripts/init-local.sh new file mode 100755 index 0000000..12244a8 --- /dev/null +++ b/scripts/init-local.sh @@ -0,0 +1,28 @@ +#!/bin/sh +set -eu +cd "$(dirname "$0")/.." +mkdir -p secrets +if [ ! -f secrets/postgres_password ]; then + umask 077 + openssl rand -hex 24 > secrets/postgres_password +fi +if [ ! -f secrets/redis_password ]; then + umask 077 + openssl rand -hex 24 > secrets/redis_password +fi +postgres_password=$(cat secrets/postgres_password) +redis_password=$(cat secrets/redis_password) +umask 077 +printf 'postgres://messenger:%s@postgres:5432/messenger?sslmode=disable\n' "$postgres_password" > secrets/database_url +printf 'redis://:%s@redis:6379/0\n' "$redis_password" > secrets/cache_url +if [ ! -f .env ]; then cp .env.example .env; fi +for secret_name in s3_admin_access s3_admin_secret s3_access_key_id s3_secret_access_key; do + if [ ! -f "secrets/$secret_name" ]; then openssl rand -hex 24 > "secrets/$secret_name"; fi +done +printf '%s\n' 'S3 credentials ready; bucket and policy are bootstrapped by s3-init.' + +if [ ! -f secrets/broker_password ]; then openssl rand -hex 24 > secrets/broker_password; fi +broker_password=$(cat secrets/broker_password) +printf 'redis://:%s@broker:6379/0\n' "$broker_password" > secrets/broker_url + +if [ ! -f secrets/grafana_admin_password ]; then openssl rand -hex 24 > secrets/grafana_admin_password; fi diff --git a/scripts/restart.sh b/scripts/restart.sh new file mode 100755 index 0000000..ed28737 --- /dev/null +++ b/scripts/restart.sh @@ -0,0 +1,28 @@ +#!/bin/sh +set -eu +cd "$(dirname "$0")/.." + +wait_healthy() { + service=$1 + container_id=$(docker compose --env-file .env ps -q "$service") + attempts=0 + until [ "$(docker inspect --format '{{.State.Health.Status}}' "$container_id" 2>/dev/null || true)" = healthy ]; do + attempts=$((attempts + 1)) + if [ "$attempts" -ge 60 ]; then + printf '%s\n' "$service did not become healthy" >&2 + exit 1 + fi + sleep 1 + done +} + +docker compose --env-file .env restart postgres redis s3 broker >/dev/null +wait_healthy postgres +wait_healthy redis +wait_healthy s3 +wait_healthy broker +docker compose --env-file .env restart api1 api2 cleanup dispatcher worker >/dev/null +wait_healthy api1 +wait_healthy api2 +docker compose --env-file .env restart nginx otel-collector prometheus loki tempo grafana >/dev/null +wait_healthy nginx diff --git a/scripts/start-api.sh b/scripts/start-api.sh new file mode 100755 index 0000000..af10135 --- /dev/null +++ b/scripts/start-api.sh @@ -0,0 +1,17 @@ +#!/bin/sh +set -eu +cd "$(dirname "$0")/.." + +docker compose --env-file .env start api1 api2 >/dev/null +for service in api1 api2; do + container_id=$(docker compose --env-file .env ps -q "$service") + attempts=0 + until [ "$(docker inspect --format '{{.State.Health.Status}}' "$container_id" 2>/dev/null || true)" = healthy ]; do + attempts=$((attempts + 1)) + if [ "$attempts" -ge 60 ]; then + printf '%s\n' "$service did not become healthy" >&2 + exit 1 + fi + sleep 1 + done +done diff --git a/scripts/stop-one.sh b/scripts/stop-one.sh new file mode 100755 index 0000000..c830aac --- /dev/null +++ b/scripts/stop-one.sh @@ -0,0 +1,12 @@ +#!/bin/sh +set -eu +cd "$(dirname "$0")/.." +: "${TARGET_INSTANCE_ID:?test provides target API instance alias}" + +case "$TARGET_INSTANCE_ID" in + api-1) service=api1 ;; + api-2) service=api2 ;; + *) printf '%s\n' "Unknown API instance alias" >&2; exit 2 ;; +esac + +docker compose --env-file .env stop --timeout 3 "$service" >/dev/null diff --git a/student-tests/obs_helpers.py b/student-tests/obs_helpers.py new file mode 100644 index 0000000..11e7371 --- /dev/null +++ b/student-tests/obs_helpers.py @@ -0,0 +1,65 @@ +"""Read stored signals through internal backend APIs; no public admin ports.""" +import json +import time +import urllib.parse +import urllib.request +import urllib.error +import base64 +from pipeline_helpers import * + + +def backend(url, text=False): + # Query via Grafana proxy: observation must work even while worker is stopped. + target=urllib.parse.urlsplit(url) + sources={'prometheus:9090':'prometheus','loki:3100':'loki','tempo:3200':'tempo'} + check(target.netloc in sources,'unknown telemetry backend') + password=(ROOT/'secrets/grafana_admin_password').read_text().strip() + base=os.environ.get('GRAFANA_URL','http://localhost:13000').rstrip('/') + address=base+'/api/datasources/proxy/uid/'+sources[target.netloc]+target.path + if target.query:address+='?'+target.query + auth=base64.b64encode(('admin:'+password).encode()).decode() + request=urllib.request.Request(address,headers={'Accept':'application/json','Authorization':'Basic '+auth}) + try: + with urllib.request.urlopen(request,timeout=8) as response:output=response.read().decode() + return output if text else json.loads(output) + except (urllib.error.URLError,OSError,ValueError): + raise Failure('telemetry backend not ready: '+sources[target.netloc]) from None + + +def prom(expr): + result=backend('http://prometheus:9090/api/v1/query?'+urllib.parse.urlencode({'query':expr})) + check(result['status']=='success','Prometheus query failed') + return result['data']['result'] + + +def number(expr, default=0): + values=prom(expr) + return float(values[0]['value'][1]) if values else default + + +def trace_data(trace_id): + return backend('http://tempo:3200/api/traces/'+trace_id) + + +def spans(value): + found=[] + if isinstance(value,dict): + if 'spanId' in value and 'name' in value:found.append(value) + for child in value.values():found.extend(spans(child)) + elif isinstance(value,list): + for child in value:found.extend(spans(child)) + return found + + +def logs(trace_id=None, service=None): + selector='{service_name="'+service+'"}' if service else '{service_name=~"messenger-api|messenger-worker|messenger-dispatcher"}' + query=selector+' | json' + if trace_id:query+=' | trace_id="'+trace_id+'"' + data=backend('http://loki:3100/loki/api/v1/query_range?'+urllib.parse.urlencode({'query':query,'start':str(time.time_ns()-900_000_000_000),'limit':'1000'})) + return [json.loads(row[1]) for stream in data['data']['result'] for row in stream['values']] + + +def alert(name): + result=backend('http://prometheus:9090/api/v1/alerts') + matches=[a for a in result['data']['alerts'] if a['labels'].get('alertname')==name] + return matches[0]['state'] if matches else 'resolved' diff --git a/student-tests/pipeline_helpers.py b/student-tests/pipeline_helpers.py new file mode 100644 index 0000000..7446ce6 --- /dev/null +++ b/student-tests/pipeline_helpers.py @@ -0,0 +1,82 @@ +"""Local synthetic tests; credentials stay in memory and never enter evidence.""" +import json +import os +from pathlib import Path +import subprocess +import sys +import time +ROOT = Path(__file__).resolve().parents[1] +sys.path.insert(0, str(ROOT / 'tests')) +from client import Client, Failure, check, run + + +def compose(*args, data=None): + result = subprocess.run(['docker', 'compose', '--env-file', '.env', *args], cwd=ROOT, + input=data, capture_output=True, text=True) + check(result.returncode == 0, 'compose command failed: ' + ' '.join(args[:3])) + return result.stdout.strip() + + +def psql(sql): + return compose('exec', '-T', 'postgres', 'psql', '-XAtq', '-v', 'ON_ERROR_STOP=1', + '-U', 'messenger', '-d', 'messenger', data=sql) + + +def wait_for(test, message, timeout=100): + deadline = time.monotonic() + timeout + while time.monotonic() < deadline: + try: + value = test() + if value: + return value + except Failure: + pass + time.sleep(.3) + raise Failure(message) + + +def ready(client): + wait_for(lambda: client.call('GET', '/health/ready', schema='Health'), 'API not ready') + + +def upload(client, actor, chat, raw=None): + if raw is None: + raw = (ROOT / 'tests/fixtures/sample.png').read_bytes() + result = client.call('POST', '/api/v1/chats/' + chat['id'] + '/attachments?filename=test.png', + 202, actor=actor, raw=raw, headers={'Content-Type': 'image/png'}, schema='Attachment')['json'] + check(result['status'] == 'queued', '202 must contain queued snapshot') + return result + + +def state(client, actor, attachment): + return client.call('GET', '/api/v1/attachments/' + attachment['id'], actor=actor, schema='Attachment')['json'] + + +def wait_state(client, actor, attachment, wanted='ready', timeout=100): + def poll(): + result = state(client, actor, attachment) + if wanted == 'ready': + check(result['status'] != 'failed', 'job unexpectedly failed') + return result if result['status'] == wanted else None + return wait_for(poll, 'attachment did not become ' + wanted, timeout) + + +def fault(point): + compose('exec', '-T', 'api1', '/messenger', 'fault', 'arm', point) + + +def hit(point): + wait_for(lambda: psql("SELECT hit_at IS NOT NULL FROM pipeline_faults WHERE point='" + point + "';") == 't', + 'fault not reached: ' + point) + + +def dump_evidence(name, value): + (ROOT / 'evidence').mkdir(exist_ok=True) + (ROOT / 'evidence' / name).write_text(json.dumps(value, ensure_ascii=False, indent=2) + '\n') + + +def start_only(service): + # Compose start can also start dependencies; fault tests need exactly this container. + container = compose('ps', '--all', '-q', service) + result = subprocess.run(['docker', 'start', container], capture_output=True, text=True) + check(result.returncode == 0, 'cannot start container ' + service) diff --git a/student-tests/verify_incidents.py b/student-tests/verify_incidents.py new file mode 100644 index 0000000..a1881b0 --- /dev/null +++ b/student-tests/verify_incidents.py @@ -0,0 +1,129 @@ +"""Two diagnosed incidents + bounded telemetry outage, using stored observations.""" +import concurrent.futures +import datetime +import json +import os +import subprocess +import time +import uuid +from obs_helpers import * + + +def send_load(client,actor,chat,seconds,workers=2): + deadline=time.monotonic()+seconds + path='/api/v1/chats/'+chat['id']+'/messages' + def task(_): + c=Client(client.base,timeout=8);ok=failed=0;durations=[] + while time.monotonic()=3 and queue['oldest_age_seconds']>10,'queue metrics do not explain lag') + check(queue['worker_heartbeat_age_seconds']>10,'worker absence not observable') + client.message(actor,chat,'messages continue with worker stopped') + start_only('worker') + for item in pending:wait_state(client,actor,item) + wait_for(lambda:alert('MessengerQueueLag')=='resolved','queue alert did not resolve',40) + queue.update(resolved=True,depth_after=number('max(messenger_queue_depth)')) + output['worker_stopped']=queue + print('PASS: queue lag firing -> resolved, heartbeat stale, backlog drained',flush=True) + + # 2. PostgreSQL unavailable; identify dependency spans, not generic slow handler. + compose('stop','postgres');begin=time.monotonic();error_trace=None;errors=0 + while time.monotonic()-begin<50: + response=client.call('POST','/api/v1/chats/'+chat['id']+'/messages',503,actor=actor, + body={'text':'synthetic outage'},headers={'Idempotency-Key':uuid.uuid4().hex}) + error_trace=response['headers']['X-Trace-Id'];errors+=1 + if alert('MessengerHighErrorRate')=='firing' and time.monotonic()-begin>=15:break + check(alert('MessengerHighErrorRate')=='firing','5xx alert did not fire') + database=dict(firing=True,detected_seconds=round(time.monotonic()-begin,3),failed_requests=errors, + error_ratio=number('messenger:api_error_ratio:rate1m'), + db_available=number('min_over_time(messenger_db_available[1m])'), + pool_waiting_peak=number('max(max_over_time(messenger_db_pool_waiting[1m]))'),trace_id=error_trace) + data=wait_for(lambda:trace_data(error_trace),'DB failure trace not stored',25) + children=[s for s in spans(data) if s['name'].startswith('db.')] + check(children,'no DB child span explains HTTP failure') + check(any(s.get('status',{}).get('code') in (2,'STATUS_CODE_ERROR') for s in children),'DB spans lack error status') + check(database['db_available']==0,'DB availability signal did not show outage') + dump_evidence('db-incident-trace.json',data) + compose('start','postgres');ready(client) + # Real recovered traffic, no time-series manipulation. + deadline=time.monotonic()+90 + while alert('MessengerHighErrorRate')!='resolved' and time.monotonic() resolved',flush=True) + + # 3. A telemetry backend down: business writes still work; buffers are finite. + before_failed=number('sum(otelcol_exporter_send_failed_spans_total{exporter="otlphttp/tempo"})') + before_stats=collector_stats() + since=datetime.datetime.now(datetime.timezone.utc).isoformat() + compose('stop','tempo') + result=send_load(client,actor,chat,20) + check(result['successful_requests']>0 and result['failed_requests']==0,'telemetry backend broke business API') + # wait beyond bounded retry deadline so final drops can be observed. + time.sleep(8) + failed=number('sum(otelcol_exporter_send_failed_spans_total{exporter="otlphttp/tempo"})')-before_failed + queued=number('max(max_over_time(otelcol_exporter_queue_size{exporter="otlphttp/tempo"}[1m]))') + check(failed>0,'Collector failures not observable') + check(queued<=32,'Collector exceeded configured buffer') + after_stats=collector_stats() + raw_logs=compose('logs','--no-log-prefix','--since',since,'otel-collector') + dropped=0 + for line in raw_logs.splitlines(): + try:record=json.loads(line) + except ValueError:continue + if record.get('otelcol.signal')=='traces':dropped+=int(record.get('dropped_items',0)) + check(dropped>0,'actual dropped trace records not counted') + output['tempo_stopped']=dict(workload=result,collector_before=before_stats,collector_during=after_stats, + failed_span_send_records_delta=failed,max_queue_batches=queued,queue_limit_batches=32, + dropped_spans_from_collector_logs=dropped,business_api_preserved=True) + compose('start','tempo') + wait_for(lambda:backend('http://tempo:3200/ready',text=True),'Tempo did not recover',40) + probe=client.call('GET','/api/v1/users/me',actor=actor) + wait_for(lambda:trace_data(probe['headers']['X-Trace-Id']),'traces did not resume',40) + output['tempo_stopped']['export_resumed']=True + print('PASS: telemetry backend outage preserves writes, queues bounded, dropped spans counted',flush=True) + dump_evidence('incidents-grade4.json',output) + finally: + compose('start','postgres','worker','tempo');ready(client) + + +if __name__=='__main__': + run(main) diff --git a/student-tests/verify_observability.py b/student-tests/verify_observability.py new file mode 100644 index 0000000..c01119c --- /dev/null +++ b/student-tests/verify_observability.py @@ -0,0 +1,94 @@ +"""Grade 4: real stored signals, W3C chain with retry, dashboard and privacy.""" +import base64 +import hashlib +import os +import time +import uuid +from obs_helpers import * + + +def normalized_id(value): + if not value:return '' + if all(c in '0123456789abcdef' for c in value.lower()) and len(value) in (16,32):return value.lower() + return base64.b64decode(value).hex() + + +def main(): + client=Client(timeout=30) + actor,_,_,chat=client.bootstrap() + marker='PRIVATE-BODY-'+uuid.uuid4().hex + client.message(actor,chat,marker) + trace_id=uuid.uuid4().hex + raw=(ROOT/'tests/fixtures/sample.png').read_bytes() + try: + compose('stop','worker') + response=client.call('POST','/api/v1/chats/'+chat['id']+'/attachments?filename=private-canary.png',202, + actor=actor,raw=raw,headers={'Content-Type':'image/png','traceparent':'00-'+trace_id+'-0123456789abcdef-01','tracestate':'lab=synthetic'},schema='Attachment') + item=response['json'];request_id=response['headers']['X-Request-Id'] + check(response['headers']['X-Trace-Id']==trace_id,'incoming W3C traceparent not honored') + compose('stop','s3');start_only('worker') + wait_for(lambda: psql("SELECT attempts>=1 AND state='queued' FROM attachment_jobs WHERE id='"+item['id']+"';")=='t','retry not observed',35) + compose('start','s3');ready(client) + wait_state(client,actor,item) + check(client.call('GET','/api/v1/attachments/'+item['id']+'/content',actor=actor)['bytes']==raw,'original changed') + def linked(): + data=trace_data(trace_id);found=spans(data) + names={s['name'] for s in found} + consumers=[s for s in found if s['name']=='attachment.consume'] + if len(consumers)>=2 and {'attachment.publish','image.process','s3.put','db.update','db.commit'}.issubset(names):return data + return None + data=wait_for(linked,'complete upload/publish/retry/consume/result trace missing',60) + found=spans(data) + publishers={normalized_id(s['spanId']) for s in found if s['name']=='attachment.publish'} + consumers=[s for s in found if s['name']=='attachment.consume'] + check(all(normalized_id(s.get('parentSpanId','')) in publishers for s in consumers),'consumer parent is not producer') + correlated=wait_for(lambda: [entry for entry in logs(trace_id) if entry.get('event')=='job.completed' and entry.get('outcome')=='success'], 'worker log missing',40) + all_logs=logs(trace_id) + events={entry['event'] for entry in all_logs} + check({'http.request','job.published','job.started','job.completed'}.issubset(events),'API/publisher/worker logs not correlated') + check(any(e.get('request_id')==request_id and e['event']=='http.request' for e in all_logs),'request id correlation missing') + check(correlated[0]['request_id']==request_id,'worker lost originating request id') + encoded=json.dumps(data)+json.dumps(all_logs)+json.dumps(logs(service='messenger-api')) + for secret in (marker,actor['token'],actor['registration']['password'],'private-canary.png'): + check(secret not in encoded,'sensitive payload/header/filename entered telemetry') + # Actual stored metric names, not assumptions from SDK naming. + expressions={ + 'http_requests':'sum(messenger_http_requests_total)', + 'http_histogram':'sum(messenger_http_duration_seconds_count)', + 'jobs_success':'sum(messenger_jobs_processed_total{outcome="success"})', + 'jobs_retry':'sum(messenger_jobs_processed_total{outcome="retry"})', + 'ready_histogram':'sum(messenger_time_to_ready_seconds_count)', + 'queue':'max(messenger_queue_depth)', + 'pool':'max(messenger_db_pool_capacity)', + 'worker_capacity':'max(messenger_worker_capacity)', + 'memory':'max(messenger_container_memory_bytes)', + } + values={} + for name,expression in expressions.items(): + wait_for(lambda: prom(expression),'missing metric '+name,20) + values[name]=number(expression) + check(values['jobs_success']>0 and values['jobs_retry']>0,'worker outcomes absent') + metric_rows=prom('{__name__=~"messenger_.*"}') + forbidden={'user_id','chat_id','message_id','attachment_id','request_id','filename','url','path','job_id'} + check(not any(forbidden.intersection(row['metric']) for row in metric_rows),'high-cardinality metric labels') + password=(ROOT/'secrets/grafana_admin_password').read_text().strip() + code=''' +import base64,json,sys,urllib.request +p=sys.stdin.read().strip();req=urllib.request.Request('http://grafana:3000/api/dashboards/uid/mtusi-messenger',headers={'Authorization':'Basic '+base64.b64encode(('admin:'+p).encode()).decode()}) +with urllib.request.urlopen(req,timeout=10) as r:d=json.load(r) +print(json.dumps({'title':d['dashboard']['title'],'panels':len(d['dashboard']['panels'])})) +''' + dashboard=json.loads(compose('exec','-T','worker','python3','-c',code,data=password)) + check(dashboard['panels']>=18,'provisioned dashboard missing') + dump_evidence('trace-example.json',data) + dump_evidence('correlated-logs.json',all_logs) + output=dict(status='PASS',trace_id=trace_id,request_id=request_id,consumer_attempt_spans=len(consumers), + events=sorted(events),metrics=values,dashboard=dashboard,privacy='PASS',low_cardinality='PASS') + dump_evidence('observability-grade4.json',output) + print('PASS:',json.dumps(output)) + finally: + compose('start','s3','worker');ready(client) + + +if __name__=='__main__': + run(main) diff --git a/student-tests/verify_pipeline.py b/student-tests/verify_pipeline.py new file mode 100644 index 0000000..cb2c37a --- /dev/null +++ b/student-tests/verify_pipeline.py @@ -0,0 +1,138 @@ +"""Grade 4: durable acceptance, outbox, crashes, lease, retries, explicit replay.""" +import hashlib +import json +import time +import uuid +from pipeline_helpers import * + + +def main(): + client = Client(timeout=30) + ready(client) + actor, _, _, chat = client.bootstrap() + raw = (ROOT / 'tests/fixtures/sample.png').read_bytes() + results = {} + try: + # Worker absent: no resize in API, accepted job survives real restarts. + compose('stop', 'worker') + attachment = upload(client, actor, chat) + client.call('GET', '/api/v1/attachments/' + attachment['id'] + '/content', 409, actor=actor) + client.call('POST', '/api/v1/chats/' + chat['id'] + '/messages', 201, actor=actor, + body={'text': '', 'attachment_ids': [attachment['id']]}, + headers={'Idempotency-Key': uuid.uuid4().hex}, schema='Message') + services = ('api1', 'api2', 'broker') + def starts(): + values = {} + for service in services: + container = compose('ps', '-q', service) + data = subprocess.run(['docker', 'inspect', container], capture_output=True, text=True, check=True) + values[service] = json.loads(data.stdout)[0]['State']['StartedAt'] + return values + before = starts() + compose('restart', *services) + ready(client) + after = starts() + check(all(before[s] != after[s] for s in services), 'actual restart not observed') + check(state(client, actor, attachment)['status'] == 'queued', 'worker stopped but state changed') + start_only('worker') + result = wait_state(client, actor, attachment) + check({v['name'] for v in result['variants']} == {'thumbnail', 'crop', 'optimized'}, 'missing grade4 variants') + check(client.call('GET', '/api/v1/attachments/' + attachment['id'] + '/content', actor=actor)['bytes'] == raw, 'original changed') + wait_for(lambda: psql("SELECT staging_deleted_at IS NOT NULL FROM attachment_jobs WHERE id='" + attachment['id'] + "';") == 't', 'staging not cleaned') + results['stopped_worker_and_restart'] = dict(status='PASS', before=before, after=after) + + # Committed outbox remains when dispatcher dies before XADD. + compose('stop', 'dispatcher') + fault('dispatcher_before_publish') + item = upload(client, actor, chat) + check(psql("SELECT count(*) FROM attachment_outbox WHERE job_id='" + item['id'] + "';") == '1', 'outbox absent') + compose('start', 'dispatcher') + hit('dispatcher_before_publish') + wait_state(client, actor, item) + results['commit_publish_gap'] = 'PASS: actual dispatcher exit 86 and automatic restart' + + # Crash after S3 objects, before final PG COMMIT; reclaim expired processing. + fault('worker_after_objects') + item = upload(client, actor, chat) + hit('worker_after_objects') + check(psql("SELECT state FROM attachment_jobs WHERE id='" + item['id'] + "';") == 'processing', 'crash did not leave processing') + started = time.monotonic() + result = wait_state(client, actor, item, timeout=110) + check(psql("SELECT attempts FROM attachment_jobs WHERE id='" + item['id'] + "';") == '2', 'expired lease not recovered') + check(len(result['variants']) == 3, 'duplicate variant metadata') + results['crash_after_objects'] = dict(status='PASS', recovery_seconds=round(time.monotonic()-started, 3), attempts=2) + + # COMMIT ready then death before ack: ready must never regress/reprocess. + fault('worker_before_ack') + item = upload(client, actor, chat) + hit('worker_before_ack') + result = wait_state(client, actor, item) + attempts = psql("SELECT attempts FROM attachment_jobs WHERE id='" + item['id'] + "';") + duplicate = dict(schema_version=1, job_id=item['id'], type='attachment.process.v1', attachment_id=item['id'], + pipeline_version=1, attempt=1, created_at=item['created_at']) + # Use broker password only inside the container, never in evidence. + duplicate_id = compose('exec', '-T', 'broker', 'sh', '-ec', + 'redis-cli --no-auth-warning -a "$(cat /run/secrets/broker_password)" XADD attachment.jobs.v1 "*" job "$1"', + 'sh', json.dumps(duplicate)) + wait_for(lambda: compose('exec', '-T', 'broker', 'sh', '-ec', + 'redis-cli --raw --no-auth-warning -a "$(cat /run/secrets/broker_password)" XRANGE attachment.jobs.v1 "$1" "$1"', + 'sh', duplicate_id) == '', 'ready duplicate was not acknowledged') + check(state(client, actor, item) == result, 'ready duplicate changed result') + check(psql("SELECT attempts FROM attachment_jobs WHERE id='" + item['id'] + "';") == attempts, 'ready duplicate processed again') + results['commit_before_ack_and_duplicate'] = 'PASS' + + # Temporary S3 failure: attempt budget/backoff, failed registry, audited replay. + compose('stop', 'worker') + item = upload(client, actor, chat) + psql("UPDATE attachment_jobs SET max_attempts=2 WHERE id='" + item['id'] + "';") + compose('stop', 's3') + start_only('worker') + check(compose('ps', '--status', 'running', '-q', 's3') == '', 'fault test unexpectedly restarted S3') + wait_for(lambda: psql("SELECT state FROM attachment_jobs WHERE id='" + item['id'] + "';") == 'failed', 'temporary errors did not exhaust retries', 100) + compose('start', 's3') + ready(client) + failed = state(client, actor, item) + check(failed['error_code'] == 'retry_exhausted', 'unsafe or missing final error') + client.call('GET', '/api/v1/attachments/' + item['id'] + '/content', 409, actor=actor) + client.call('POST', '/api/v1/chats/' + chat['id'] + '/messages', 409, actor=actor, + body={'text':'failed', 'attachment_ids':[item['id']]}, headers={'Idempotency-Key':uuid.uuid4().hex}) + registry = compose('exec', '-T', 'dispatcher', '/messenger', 'failed-jobs') + check(item['id'] in registry, 'failed registry missing job') + compose('exec', '-T', 'dispatcher', '/messenger', 'replay', item['id']) + wait_state(client, actor, item) + check(psql("SELECT count(*) FROM attachment_replays WHERE job_id='" + item['id'] + "';") == '1', 'replay audit missing') + proc = subprocess.run(['docker','compose','--env-file','.env','exec','-T','dispatcher','/messenger','replay',item['id']], cwd=ROOT, capture_output=True) + check(proc.returncode != 0, 'ready replay must be refused') + results['retry_failed_registry_replay'] = dict(status='PASS', exhausted_attempts=2, error_code='retry_exhausted', replay_audit_rows=1, ready_replay_rejected=True) + + # Permanent invalid staging: no endless retries, safe machine error. + compose('stop', 'worker') + item = upload(client, actor, chat) + original_key = uuid.uuid4().hex + original_body = {'text':'accepted while queued', 'attachment_ids':[item['id']]} + original_message = client.call('POST', '/api/v1/chats/' + chat['id'] + '/messages', 201, + actor=actor, body=original_body, headers={'Idempotency-Key':original_key}, schema='Message')['json'] + # Integrity metadata mismatch represents invalid durable source. + psql("UPDATE attachments SET sha256=repeat('0',64) WHERE id='" + item['id'] + "';") + start_only('worker') + failed = wait_state(client, actor, item, 'failed') + check(failed['error_code'] == 'invalid_image', 'permanent invalid source must fail safely') + check(psql("SELECT attempts FROM attachment_jobs WHERE id='" + item['id'] + "';") == '1', 'permanent error retried') + repeated = client.call('POST', '/api/v1/chats/' + chat['id'] + '/messages', 201, + actor=actor, body=original_body, headers={'Idempotency-Key':original_key}, schema='Message')['json'] + check(repeated == original_message, 'committed message replay changed after attachment failed') + client.call('POST', '/api/v1/chats/' + chat['id'] + '/messages', 409, + actor=actor, body=original_body, headers={'Idempotency-Key':uuid.uuid4().hex}) + results['permanent_error'] = 'PASS' + results['message_replay_after_attachment_failure'] = 'PASS' + results['crash_points'] = json.loads(psql("SELECT json_agg(json_build_object('point',point,'armed',armed,'hit_at',hit_at) ORDER BY point) FROM pipeline_faults WHERE hit_at IS NOT NULL;")) + dump_evidence('pipeline-grade4.json', results) + print('PASS:', json.dumps(results)) + finally: + compose('start', 's3', 'broker', 'dispatcher', 'worker', 'api1', 'api2') + psql('DELETE FROM pipeline_faults;') + ready(client) + + +if __name__ == '__main__': + run(main) diff --git a/student-tests/verify_resilience.py b/student-tests/verify_resilience.py new file mode 100644 index 0000000..1de8eda --- /dev/null +++ b/student-tests/verify_resilience.py @@ -0,0 +1,92 @@ +"""Public persistence/failover hooks plus actual container restart evidence.""" +import json +import datetime +import subprocess +import uuid +from pipeline_helpers import * + + +def snapshots(): + services=('postgres','redis','s3','broker','api1','api2','dispatcher','worker','cleanup','nginx','otel-collector','prometheus','loki','tempo','grafana') + result={} + for service in services: + container=compose('ps','-q',service) + info=subprocess.run(['docker','inspect',container],capture_output=True,text=True,check=True) + data=json.loads(info.stdout)[0] + result[service]=dict(id=data['Id'],started_at=data['State']['StartedAt']) + return result + + +def main(): + client=Client();ready(client) + before=snapshots() + base=client.base + command=['make','test-persistence','ACTION_SCRIPT=scripts/restart.sh','BASE_URL='+base] + result=subprocess.run(command,cwd=ROOT,capture_output=True,text=True) + check(result.returncode==0,'public persistence failed: '+result.stdout[-500:]) + after=snapshots() + check(all(before[s]['started_at']!=after[s]['started_at'] for s in before),'hook did not restart all required services') + + output={'persistence':dict(status='PASS',before=before,after=after),'failover':[]} + public=subprocess.run(['make','test-failover','ACTION_SCRIPT=scripts/stop-one.sh','BASE_URL='+base],cwd=ROOT,capture_output=True,text=True) + check(public.returncode==0,'public failover failed: '+public.stdout[-500:]) + output['public_failover']=dict(status='PASS',output=public.stdout.strip()) + try: + compose('start','api1','api2');ready(client) + actor,_,_,chat=client.bootstrap() + item=upload(client,actor,chat);wait_state(client,actor,item) + raw=(ROOT/'tests/fixtures/sample.png').read_bytes() + path='/api/v1/chats/'+chat['id']+'/messages' + for target in ('api-1','api-2'): + compose('start','api1','api2');ready(client) + seen=set() + wait_for(lambda: (seen.add(client.call('GET','/health/ready')['headers']['X-Instance-Id']) or len(seen)==2),'two replicas not visible') + process=subprocess.Popen([str(ROOT/'scripts/stop-one.sh')],cwd=ROOT, + env=dict(os.environ,TARGET_INSTANCE_ID=target),stdout=subprocess.DEVNULL,stderr=subprocess.DEVNULL) + began=time.monotonic();deadline=began+8;first=None;confirmed=set();failures=0 + key=uuid.uuid4().hex + while time.monotonic()0,'metrics unavailable after restart') + stamp=datetime.datetime.fromisoformat(after['prometheus']['started_at'].replace('Z','+00:00')).timestamp()-10 + historical=number('sum(messenger_http_requests_total @ '+str(stamp)+')') + check(historical>0,'historical metrics lost after Prometheus restart') + output['metric_history_before_restart']={'unix_time':stamp,'http_counter_sum':historical} + output['telemetry_history_preserved']=True + dump_evidence('resilience.json',output) + print('PASS: all fifteen core/telemetry services restarted, same token/data/original preserved; both API failures covered') + + +if __name__=='__main__': + run(main) diff --git a/student-tests/verify_variants.py b/student-tests/verify_variants.py new file mode 100644 index 0000000..3ed28c5 --- /dev/null +++ b/student-tests/verify_variants.py @@ -0,0 +1,100 @@ +"""Grade 4 image contract: actual decoding, EXIF, alpha, limits and private storage.""" +import base64 +import hashlib +import json +import struct +import uuid +import zlib +from pipeline_helpers import * + + +def make_fixture(kind): + code = ''' +import base64,io,sys +from PIL import Image +kind=sys.argv[1] +if kind=='orientation': + im=Image.new('RGB',(320,200),'red');exif=Image.Exif();exif[274]=6;exif[315]='synthetic test';kw=dict(format='JPEG',exif=exif) +elif kind=='alpha': + im=Image.new('RGBA',(8,4),(255,0,0,0));kw=dict(format='PNG') +elif kind=='boundary': + im=Image.new('RGB',(8000,5000),'blue');kw=dict(format='PNG') +else: + im=Image.new('RGB',(1601,801),'blue');kw=dict(format='PNG') +b=io.BytesIO();im.save(b,**kw);print(base64.b64encode(b.getvalue()).decode()) +''' + return base64.b64decode(compose('exec', '-T', 'worker', 'python3', '-c', code, kind)) + + +def inspect_image(raw): + code = ''' +import base64,io,json,sys +from PIL import Image +im=Image.open(io.BytesIO(base64.b64decode(sys.stdin.read())));im.load() +print(json.dumps(dict(format=im.format,width=im.width,height=im.height,exif=len(im.getexif()),pixel=list(im.convert('RGB').getpixel((0,0)))))) +''' + return json.loads(compose('exec', '-T', 'worker', 'python3', '-c', code, data=base64.b64encode(raw).decode())) + + +def chunk(kind, payload): + return struct.pack('>I', len(payload)) + kind + payload + struct.pack('>I', zlib.crc32(kind+payload) & 0xffffffff) + + +def main(): + client=Client(timeout=60) + actor, _, outsider, chat=client.bootstrap() + results={} + for kind in ('orientation','alpha','large','boundary'): + raw=make_fixture(kind) + content_type='image/jpeg' if kind=='orientation' else 'image/png' + item=client.call('POST','/api/v1/chats/'+chat['id']+'/attachments?filename=synthetic',202,actor=actor,raw=raw, + headers={'Content-Type':content_type},schema='Attachment')['json'] + result=wait_state(client,actor,item) + variants={v['name']:v for v in result['variants']} + expected={'orientation':(160,256),'alpha':(8,4),'large':(256,128),'boundary':(256,160)}[kind] + check((variants['thumbnail']['width'],variants['thumbnail']['height'])==expected,'thumbnail dimensions/orientation') + check((variants['crop']['width'],variants['crop']['height'])==(128,128),'crop dimensions') + optimized={'orientation':(200,320),'alpha':(8,4),'large':(1280,640),'boundary':(1280,800)}[kind] + check((variants['optimized']['width'],variants['optimized']['height'])==optimized,'optimized fit dimensions') + for name,meta in variants.items(): + blob=client.call('GET','/api/v1/attachments/'+item['id']+'/content?variant='+name,actor=actor)['bytes'] + check(hashlib.sha256(blob).hexdigest()==meta['sha256'] and len(blob)==meta['size_bytes'],'variant checksum/size') + decoded=inspect_image(blob) + check(decoded['format']==('WEBP' if name=='optimized' else 'JPEG'),'actual variant encoding') + check(decoded['exif']==0,'EXIF leaked into variant') + if kind=='alpha': + check(all(v>=250 for v in decoded['pixel']),'alpha not composited on white') + check(client.call('GET','/api/v1/attachments/'+item['id']+'/content',actor=actor)['bytes']==raw,'original changed') + client.call('GET','/api/v1/attachments/'+item['id'],404,actor=outsider) + client.call('GET','/api/v1/attachments/'+item['id']+'/content?variant=crop',404,actor=outsider) + client.call('GET','/api/v1/attachments/'+item['id']+'/content?variant=missing',404,actor=actor) + results[kind]={'status':'PASS','variants':variants} + original=(ROOT/'tests/fixtures/sample.png').read_bytes() + for size in (6294458,10485760): + extra=size-len(original)-12 + raw=original[:-12]+chunk(b'paDd',bytes(extra))+original[-12:] + check(len(raw)==size,'fixture exact size') + item=upload(client,actor,chat,raw) + wait_state(client,actor,item) + check(client.call('GET','/api/v1/attachments/'+item['id']+'/content',actor=actor)['bytes']==raw,'large original changed') + path='/api/v1/chats/'+chat['id']+'/attachments?filename=bad.png' + client.call('POST',path,413,actor=actor,raw=bytes(10485761),headers={'Content-Type':'image/png'}) + for raw in (original[:40],original[:-12],b'not an image'): + client.call('POST',path,415,actor=actor,raw=raw,headers={'Content-Type':'image/png'}) + huge=b'\x89PNG\r\n\x1a\n'+chunk(b'IHDR',struct.pack('>IIBBBBB',20000,10000,8,2,0,0,0))+chunk(b'IEND',b'') + client.call('POST',path,413,actor=actor,raw=huge,headers={'Content-Type':'image/png'}) + results['byte_limits_corrupt_pixel_limits']='PASS' + item=upload(client,actor,chat) + wait_state(client,actor,item) + audit=compose('exec','-T','api1','/messenger','s3-audit','originals/'+item['id']) + check('PASS' in audit,'S3 credential audit failed') + results['private_bucket_limited_credentials']='PASS' + metrics=json.loads(compose('exec','-T','worker','/messenger','memory')) + check(metrics['memory.peak'] < metrics['memory.max'],'worker reached memory limit') + results['worker_memory_bytes']=metrics + dump_evidence('variants-grade4.json',results) + print('PASS: actual JPEG/WebP, EXIF orientation, white alpha, bounds, ACL, checksums, private S3') + + +if __name__=='__main__': + run(main) diff --git a/worker/process_image.py b/worker/process_image.py new file mode 100644 index 0000000..55be9a2 --- /dev/null +++ b/worker/process_image.py @@ -0,0 +1,60 @@ +"""Bounded subprocess: trusted local paths, no URLs, no credentials.""" +import hashlib +import json +import resource +import sys +import warnings +from pathlib import Path +from PIL import Image, ImageOps, UnidentifiedImageError + +# Bound decoder subprocess independently of container and Go memory. +resource.setrlimit(resource.RLIMIT_AS, (1024 * 1024 * 1024, 1024 * 1024 * 1024)) +Image.MAX_IMAGE_PIXELS = 40_000_000 +warnings.simplefilter('error', Image.DecompressionBombWarning) + + +def contain(image, bound): + w, h = image.size + if max(w, h) <= bound: + return image.copy() + if w >= h: + size = (bound, max(1, h * bound // w)) + else: + size = (max(1, w * bound // h), bound) + return image.resize(size, Image.Resampling.LANCZOS) + + +def main(): + source, target = map(Path, sys.argv[1:3]) + if source.stat().st_size > 10_485_760: + raise ValueError('size') + with Image.open(source) as original: + if original.format not in ('PNG', 'JPEG') or original.width * original.height > 40_000_000: + raise ValueError('format or pixels') + original.load() + oriented = ImageOps.exif_transpose(original) + rgba = oriented.convert('RGBA') + image = Image.new('RGB', rgba.size, 'white') + image.paste(rgba, mask=rgba.getchannel('A')) + # New pixel images + empty info ensure EXIF/ICC/text are never propagated. + variants = [] + for name, result, kind, quality in ( + ('thumbnail', contain(image, 256), 'JPEG', 85), + ('crop', ImageOps.fit(image, (128, 128), method=Image.Resampling.LANCZOS, centering=(.5, .5)), 'JPEG', 85), + ('optimized', contain(image, 1280), 'WEBP', 80), + ): + result.info.clear() + path = target / name + result.save(path, format=kind, quality=quality, **({'method': 4} if kind == 'WEBP' else {})) + data = path.read_bytes() + variants.append(dict(name=name, content_type='image/webp' if kind == 'WEBP' else 'image/jpeg', + size_bytes=len(data), width=result.width, height=result.height, + sha256=hashlib.sha256(data).hexdigest())) + print(json.dumps(variants)) + + +if __name__ == '__main__': + try: + main() + except (UnidentifiedImageError, OSError, ValueError, Image.DecompressionBombError, Image.DecompressionBombWarning): + sys.exit(2) # permanent invalid_image; details never reach public API -- 2.54.0