Files
mtusi2026lab3/evidence/rolling-restart.json
T
2026-11-04 16:11:13 +03:00

38 lines
1.1 KiB
JSON

{
"operation": "serial API replica restart under load; same image, no schema change",
"duration_seconds": 40,
"concurrency": 16,
"restarted_replicas": [
{
"instance_id": "4fd03cf9f7bc",
"ready_seconds": 6.126
},
{
"instance_id": "52406027111f",
"ready_seconds": 6.148
}
],
"load": {
"lab": 3,
"workload": "closed-loop POST messages; no automatic retries",
"requested_duration_seconds": 40.0,
"elapsed_seconds": 40.161,
"concurrency": 16,
"successful_requests": 7054,
"failed_requests": 2,
"error_ratio": 0.000283,
"successful_rps": 175.644,
"successful_latency_ms": {
"p50": 104.831,
"p95": 196.305,
"p99": 278.076
},
"note": "Latency percentiles cover successes only. Timeouts/errors are counted separately; closed-loop has coordinated omission and client overhead."
},
"load_raw_file": "rolling-restart-load.json",
"limitations": [
"This verifies graceful process replacement, not a new-image rolling deployment.",
"PostgreSQL and Redis are single instances; their failure is not covered."
]
}