Confluence Runbook Page for an Operational Task
Write a runbook with clear triggers, steps and escalation.
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
Write a runbook page for this task: [TASK_NAME]. When it is used (trigger): [TRIGGER]. Systems and access needed: [SYSTEMS]. Steps as people currently do them: [CURRENT_STEPS]. Who to contact if it goes wrong: [ESCALATION]. Format: Purpose; When to use it; Before you start (checks); Steps (numbered, one action each, with the expected result and what to do if it does not appear); Rollback or undo; Verification; Signs that you have succeeded; When to stop and call for help; Escalation table (role, name placeholder, hours of cover); Related pages; Owner and Last tested date. Write the steps so they can be followed under pressure at night: no ambiguity, no jargon, one action per step. Put a visible warning line on any step that changes data or affects customers, and give any step that deletes, overwrites or sends something a confirmation check before it and a way to undo it after. List the three things most likely to go wrong, with symptoms, in a troubleshooting section. Where my steps are ambiguous, write the question for the person who knows. Do not include passwords, tokens or personal data; use a placeholder such as [SECRET] (held in the password manager). Say who may approve exceptions and how changes are reviewed and recorded. Add an annex table recording each use: date, person, trigger time, steps completed, time taken, anything unexpected, and an escalation message template (what happened, what has been tried, what is needed). Add a quarterly drill in a safe test setting with three review questions: could a new person follow it, did any step differ from reality, and what failed. Flag a runbook unused or untested for six months for a drill. It must be tested by a person before anyone relies on it; describe only what I supplied.
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.
See an example result (fictional data)
Invented example: Task: restart the nightly report job. Trigger: the 06:00 report email has not arrived by 06:30. Steps: 1. Open the job console and find the job "NightlyReport"; expected: status shows Failed or Stuck. 2. Select Restart; expected: status changes to Running. 3. Wait 10 minutes and confirm the email arrives; if not, go to Escalation. Escalation: platform on-call, then team lead after 30 minutes. Last tested: [DATE]. Second part of the page. Verification: open the report folder and confirm that tonight's file exists and has a size above 50 KB. Rollback: not needed, because a restart does not change data. Contacts table: platform on-call (24 hours), team lead (09:00 to 17:00). Review: this runbook is tested every quarter by someone who has not used it before.
Illustrative only, made-up data. Your results will differ and must be reviewed.