Keep the full output of a failed compose up in track 1 runs

The orchestrator printed only the last 500 characters of stderr before
retrying, which hid the cause of a flaky up on eight replicas. The whole
output of each failed attempt now goes to compose-up-failed.log next to
summary.csv.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01V1d3RLHNBBMA6UqBLLJv27
This commit is contained in:
Artyom
2026-10-29 23:12:55 +03:00
co-authored by Claude Opus 5.5
parent 0ce7d6ecf0
commit 16b6b28ea9
2 changed files with 8 additions and 4 deletions
@@ -9,7 +9,7 @@
- [ ] 2.1 `Makefile`: `BASE_URL` по порту `lb`, `API_IMAGE` по каталогу, `detection` трижды, `track1-control`
- [x] 2.2 `scripts/evidence-storage.sh` — volumes своего проекта; `scripts/limits.py` — коммит в заголовке
- [ ] 2.3 `scripts/scaling.py` — полный stderr сбоя `compose up` в `compose-up-failed.log`
- [x] 2.3 `scripts/scaling.py` — полный stderr сбоя `compose up` в `compose-up-failed.log`
- [ ] 2.4 `scripts/failover-each.sh` — `docker ps -a` реплик после хука; `scripts/failover-evidence.sh` — `SHIFT`
- [ ] 2.5 Проверка группы: `make lint`, `make check`, `bash -n`/`sh -n` скриптов