Triage an incident with a safe runbook | DevPrompt Lab

A practical incident workflow that favors reversible mitigation and clear ownership over risky improvisation.

Reusable content

For the incident below, create a concise runbook. Start with user impact and affected systems. Separate confirmed facts from hypotheses. Give safe, read-only checks first, then reversible mitigation options with risks and rollback steps. Identify who owns each action, what evidence confirms recovery, and when to escalate. Do not recommend deleting data, disabling security controls, or making an irreversible change without explicit evidence and a recovery plan. End with a short status update template and follow-up items.

Service: {{SERVICE}}

Alert and timeline: {{ALERT}}

Known impact: {{IMPACT}}

Available rollback: {{ROLLBACK}}

How to use it

Use sanitized telemetry. Add service-specific commands only after validating them against the team's runbook.

Example

Mitigation: route traffic to the last healthy deployment. Verify: error rate returns to baseline for 10 minutes. Rollback: restore the recorded deployment ID.