sre-runbooklisted

# sre-runbook Produces an **operational runbook**: a *tactical* procedure scoped to one alert/condition, written for an on-call engineer acting under time pressure. It answers "this page just fired — what exactly do I do?" with copy-pasteable commands and explicit decision points. A runbook is **tactical**, not strategic. Contrast: a *playbook* is strategic — it coordinates the broader response (comms, roles, multi-team escalation) across a class of incidents. A runbook is the narrow, executable fix for one named condition. Keep strategy out; link to the playbook instead. ## The "5 A's" quality frame Every section is judged against five properties: - **Actionable** — concrete commands and decisions, never "investigate the issue". - **Accurate** — commands are correct and current; thresholds match the real SLO. - **Authoritative** — it is the single source of truth for this alert; owner named. - **Accessible** — findable, readable at 3 a.m., no tribal knowledge assumed. - **Adaptable** — versioned and dated so it tracks the system as it changes. ## Pattern (seven canonical sections, in order) 1. **Overview** — the service this covers and the one alert/condition it handles. 2. **Prerequisites & Access** — tools, credentials, roles, and dashboards needed *before* you start. Get access blockers out of the way first. 3. **Detection** — the firing alert, observable symptoms, and the dashboards/ queries that confirm it. State the SLO/threshold being breached. 4. **Diagn