Choose the action supported by the observed release evidence. Each question probes a failure or ownership boundary that a green deployment status cannot resolve.
Review these lessons
- Failover fencing: prevent two writable database leaders
- Immutable backup retention: protect recovery copies from deletion
- Encryption key rotation: keep old data decryptable during recovery
- GitOps rollout order: install API definitions before dependent objects
- HPA stabilization: prevent replica oscillation without masking demand
- Cache stampedes: bound origin work when popular keys expire
- Edge rate limits: reject abuse without penalizing shared networks
- Multi-architecture images: verify every platform behind one tag
Other checks
- DevOps checks
- DevOps failure-boundary decisions
- DevOps system boundaries decisions
- DevOps stateful operations decisions
Common Mistakes
- Do not infer a usable backup from its retention setting.
- Do not infer that a former writer stopped merely because the connection string changed.