Write an On-Call Runbook for an Incident You Have Just Resolved

You have just resolved an incident and you know that in three months someone will hit the same thing at 2am and spend an hour doing what you just did. The time to write the runbook is now, while the investigation steps are fresh and you remember the wrong turns. This structures it properly.

<context>
You are a site reliability engineer with extensive incident response experience. You know that the best time to write a runbook is immediately after resolving the incident, when the steps are still fresh and the assumptions are visible. A runbook written from memory six weeks later is missing the edge cases that were only discovered during the incident.
{INCIDENT_SUMMARY} describes the incident: what broke, what symptoms appeared, what the investigation steps were, what the root cause was, and how it was resolved.
</context>

<task>
1. Structure the runbook as: symptom identification, immediate triage steps, investigation steps, resolution steps, and verification
2. Write each step in imperative language with enough specificity that someone unfamiliar with this system can follow them
3. Identify any pre-conditions or access requirements that should be noted at the top
4. Add a brief section on related issues: similar symptoms with different root causes that should be ruled out
</task>

<output_format>
- Pre-conditions and access requirements: brief list at the top
- Runbook: five sections, imperative language, numbered steps within each section
- Related issues section: 2-3 similar symptoms with brief differentiation notes
- Tone: operational, precise, written for a tired engineer at 2am
</output_format>