Runbooks

Operational procedures: the things done to a running system, written down so they're calm and repeatable instead of improvised under pressure.

Available

• Heroku → Azure migration: initial move and the reverse-direction exit.

To write (stubs)

Create a page per procedure as the need arises. Good candidates:

• Restore from backup: point-in-time restore on Flexible Server; how to verify the restored copy before promoting it.
• Rotate secrets: SECRET_KEY, DB password, Spaces keys, GHCR PAT; the order of operations so nothing drops.
• Scale up / down: change the App Service plan or Postgres SKU; what reschedules, what causes a restart.
• Promote a hotfix: fast-path a single commit through the pipeline.
• Incident response: where logs live, how to read them, how to roll back a bad deploy (redeploy the previous image tag).

Tip

Each runbook should be copy-pasteable end to end and start with a one-line "when to use this." Clarity written down now pays off under stress later.