Reference — generated from the toolkit

Source of truth: toolkit/templates/runbook.md. Edit it there; this page is regenerated on build.

Runbook — {service name}

Overview

What the service does, its dependencies, and its place in the system.

Links to the golden-signal dashboards and the alerts that page. What each alert means.

Step-by-step checks for a reported problem. Mark each step:

🤖 agent-executable — an agent may run this to gather evidence (read-only / reversible).
👤 human-only — anything irreversible (restart in prod, failover, data change) needs a named human.

Symptom	Likely cause	Action	Linked gotcha
…	…	…	`memory/…`

Who to escalate to, when, and how. Severity thresholds and the rapid-fix path.

Backup/restore and rollback procedures, and how to verify recovery succeeded.