H.7.8 Main Recovery
Reading progress2 missing
Prerequisites0/2 complete
Articles read
1 required article is still unread. Reading it first is recommended, but not required.
Article in preparation — showing conspect notes
Main recovery restores shared source after an integration failure; repository-service and product-delivery failures have different owners.
- Choose a build-cop, platform on-call, component owner, or incident commander model from incident frequency and scale.
- Classify product, flaky-evidence, infrastructure, capacity, policy, and unknown failure before retrying.
- Define authority to pause affected admission, request revert or fast fix, route ownership, and communicate scope.
- Preserve bounded escalation and handoff; coordination must not become permanent domain maintenance.
- Measure detection, classification, ownership, recovery, repeated incidents, and blocked developer time.
