Template library / deployment

Deployment recovery

Production Rollback

Restore service first, diagnose later. This procedure assumes containerized deploys behind a health endpoint.

7 required stepsStrict sequenceProduction

Restore the known-good release

  1. Pause the deploy pipelineStop new changes from extending the incident.
  2. Identify the bad releaseConfirm the version and its change set.
  3. Check out last known-goodSelect the release that was healthy before the regression.
  4. Rebuild & redeployShip the known-good build.
  5. Verify health & metricsCheck the endpoint, errors, latency, and key flows.
  6. Notify & update the status pageCommunicate the recovery.
  7. Schedule root-cause reviewLeave investigation for the follow-up.

Keep the rollback path within reach.

Run it from Airstrip when production needs a calm response.

Download Airstrip for Mac