Reliability and operations
Design for the day after launch.
Make systems observable, updateable, recoverable and supportable across normal, degraded and exception states.
Review operational readiness
Operational foundations
Make state, failure and ownership visible.
- Health, readiness and dependency state
- Logs, metrics, traces and operational evidence
- Request, event and job correlation
- Retry, idempotency and failed-work handling
- Release, rollback, backup and restoration
- Configuration, escalation and support ownership
Failure model
Recovery is an architecture path.
Reliability work identifies which failures can retry, which require compensation, what must remain idempotent, how operators find incomplete work and which dependency or team owns the next action.
Availability targets, recovery objectives, retention and support commitments remain product- and deployment-specific.
Operational review
Trace one request through success and failure.
Bring the dependencies, state changes, evidence, rollback path and support boundaries.
Talk to an architect