test(workers): prove clock and restart safety
CI and Deploy / test (pull_request) Successful in 5m13s
CI and Deploy / deploy (pull_request) Has been skipped

This commit is contained in:
cesnimda
2026-08-15 20:18:02 +02:00
parent a08ba9b45d
commit dbf28b97ce
7 changed files with 124 additions and 36 deletions
+2 -2
View File
@@ -324,9 +324,9 @@ This queue records the highest-value work that can proceed without production cr
- **Required production verification:** one-worker canary and owner-safe metrics.
- **Status:** `IMPLEMENTED — NOT VERIFIED`.
- **Blocker:** activation blocked until policy/notification prerequisites; foundation code is not blocked.
- **Evidence:** audit JT-005 service inspection; `docs/verification/bg-001-tenant-workers.md`; real-SQLite two-owner worker suite with fake email/AI; full backend 532/532; Compose validation; isolated default-off startup/health/no-export check.
- **Evidence:** audit JT-005 service inspection; `docs/verification/bg-001-tenant-workers.md`; real-SQLite two-owner worker suite with fake email/AI, exact fixed-clock threshold, fresh-instance restart idempotency and reminder/export deduplication; Compose validation; isolated default-off startup/health/no-export check.
- **Commit:** none.
- **Remaining work:** OPS-001B persistent notification/idempotency before reminders/rules; POL-001/002 and durable AI queue before enrichment; restart/clock tests; browser result surfaces; monitored single-worker production canary. Keep all four switches false.
- **Remaining work:** activation policy prerequisites, browser result surfaces and a monitored single-worker production canary. Local restart/clock coverage is complete. Keep all four switches false.
### OPS-001A — Durable operation record and lease state machine