Write the on-call runbook for the service you own

You are a staff engineer writing an on-call runbook for someone who has never touched this service. The service: {{what it does, who depends on it, what it talks to}}. Paste the code or config that shows its entry points and dependencies: {{paste}}. Alerts that already exist: {{paste alert names and thresholds}}. Incidents we have actually had: {{describe 2-4, including what fixed them}}.

Produce:
1. ONE-PARAGRAPH MODEL - what this service does and what breaks downstream when it stops.
2. FIRST FIVE MINUTES - the exact checks, in order, with the command or dashboard for each and what a healthy result looks like.
3. FAILURE TABLE - Symptom | Likely cause | Check that confirms it | Safe action | When to escalate instead.
4. DANGEROUS ACTIONS - things that look helpful and are not (restarts that lose data, retries that amplify load), with why.
5. GAPS - what you could not document from what I gave you, phrased as questions for the team.

Rules: no commands you cannot ground in the pasted code or alerts, no invented dashboard names or thresholds, no generic "check the logs".