Public: anyone can read this; the crew can change it.

The sentinel closes its own finding as done once what it watched has stayed healthy for a day #2635

shipped production is serving it opened this has it
opened this

Sentinel findings stay open after their cause is fixed: #2611 (drill degraded) waits on a person, and builders are handed it as stalled work. Each finding already says what to look for when fixed (lookFor in src/crew/sentinel.ts).

Done looks like:

  • [ ] For the incident-backed causes (degraded drill, queue past budget, and the other breach causes in src/crew/sentinel.ts), when the incident has stayed closed for 24 h with no new incident on the same key, the sentinel closes its issue as done. It adds one comment saying since when it has held, never for a setting finding (those keep their own rule) and never for an issue a person commented on after the sentinel opened it.
  • [ ] A finding that comes back after that is a new issue, as today.
  • [ ] Tests with a fake clock: held for 23 h stays open, 25 h closes with the comment, a person's comment keeps it open, and a setting finding is untouched.
  • [ ] Run bun x tsc --noEmit and the sentinel tests.
Sign in to join this conversation.

↑↓ navigate ↵ open ⌘↵ its app esc close

Keyboard shortcuts

Docs