🚨Incident Readiness Mini-Audit
cloud
Twelve questions to test whether your team can detect, respond, and learn from production incidents.
Reliability is not a dashboard theme. It is whether humans know what to do when something breaks at 2 a.m.
## How to use
Answer Yes / Partial / No. Any “No” on severity routing or ownership is a priority fix.
## Questions
1. Do we have a single place where alerts page a human?
2. Is severity defined in writing (SEV1 vs SEV3)?
3. Can we roll back the last deploy in under a documented time?
4. Do we know who owns each critical service?
5. Are runbooks linked from the alert, not buried in a wiki maze?
6. Do we test restore from backups on a schedule?
7. Is status communication assigned (not “someone will tweet”)?
8. Do we capture timelines during the incident?
9. Is there a blameless review within a fixed window after SEV1/2?
10. Are action items tracked to completion?
11. Do dependencies (auth, payments, DNS) have clear escalation paths?
12. Have we practiced at least one game day this quarter?
## Scoring
- 10–12 Yes: maintain and drill
- 6–9: fix ownership and rollback first
- Below 6: reliability work before feature velocity narratives