Incident runbook
Drafts an on-call runbook for a specific failure mode — detect, diagnose, mitigate, recover.
Saves the 3am scramble of "what do I do?".
The prompt
Copy this into ChatGPT, Claude, Gemini or any assistant you already use — then paste your own failure mode underneath it.
You are an SRE writing an on-call runbook. Given a failure mode, produce: **Title + 1-line summary** **Severity** (SEV1-SEV4) + criteria **Symptoms** (what alerts / what users see) **Diagnosis** (commands to run, dashboards to check, in order) **Mitigation** (the safest fast fix) **Rollback** (how to undo) **Long-term** (what to fix this week so it doesn't happen again) **Comms template** (1-line status-page message) Rules: - Use exact commands where applicable. - No "investigate further" — say what to look for.
Example input
Our Postgres database is at 95% disk and growing 2GB/day.
Or run it here
Clicking copies the prompt to your clipboard and opens the chatbot in a new tab. Gemini doesn't accept URL params — paste manually with Ctrl/Cmd+V.
Related prompts
Generates a multi-stage, small-image Dockerfile for a given stack.
A complete CI pipeline YAML — lint, test, build, scan, deploy — for a Node/Python/Go service.
Writes a log query (Datadog / Splunk / Loki / CloudWatch) to surface what you need.
Plan for metrics, logs, traces, and alerts for a new service — what to instrument from day 1.
Decodes any cron schedule string into when it actually runs.
Skeleton for a Terraform module — main, variables, outputs, README.
Browse the full free AI prompt library or see more devops & sre prompts.