From the House of Trae Operations Runbook. Full document: docs/HoT_Operations_Runbook.pdf.
| Level | Examples | Target response |
|---|---|---|
| P1 — Critical | Keycloak unavailable, DNS outage, reverse proxy failure, storage failure | 15 minutes |
| P2 — Major | Application outage, monitoring failure, backup failure | 1 hour |
| P3 — Minor | Dashboard issue, reporting issue, cosmetic issue | Next maintenance window |
When multiple things are down at once, restore in this order — each layer depends on the ones above it:
Infrastructure failure:
Identity failure:
Service failure: