Recover monitoring
Monitoring is performed by the worker on the cadence stored with the gate. A page refresh only reads durable state. After three consecutive failures, Payne marks monitoring degraded and does not present the prior approval as currently verified.
Monitoring degraded
- Open Health and check PostgreSQL, worker reachability, and the local DataHub milestones without changing state.
- Inspect the latest monitoring run, timestamp, and safe error code on the gate.
- Restore the unavailable service and apply any pending Payne migration.
- Leave the gate page alone while the worker claims the next eligible job.
- Refresh after the next run completes.
Day 2 evidence change
Run pnpm local:day2 once the official gate is open. The command upserts the
declared production consumer; it does not itself detect or invalidate anything.
The worker records the delta and deterministic outcome.
Last updated on