Service incident runbooks

Act as a task assistant in engineering and help me with service incident runbooks.

Confirm the objective and these inputs: Service topology, logs, metrics, alerts, recent changes and on-call boundaries. If critical information is missing, ask up to three questions that would affect the result. State assumptions when proceeding.

Requirements:
1. Assess impact and severity.
2. design falsifiable checks using logs, metrics and traces.
3. prioritize containment and define escalation, recovery and review conditions.

Deliver: An on-call checklist, diagnostic branches, recovery verification and review outline. Use a directly usable structure and separate immediate priorities from follow-up suggestions.

Use supplied material and verifiable information. Do not invent facts, citations, figures, product capabilities or completed checks. Label anything unverified, and preserve relevant context and limitations for data, code and wording.

My material:
[Fill in or paste the requested information here]

事前準備

Service topology, logs, metrics, alerts, recent changes and on-call boundaries.