Fail over an application using the DR runbook

Practice the sequence you already learned for “Fail over an application using the DR runbook” until the order feels automatic. Drill the ISO 22301 / NIST CP failover order from reliability training. DNS is not cut until the standby is healthy and the RPO is confirmed.

1. Declare the disaster with the authority named in the plan.
2. Confirm the last good recovery point.
3. Confirm the standby health checks.
4. Promote the standby database or region as written.
5. Switch traffic only after those checks pass.
6. Notify customers on the status page with the next update time.
7. Do not fail back until a separate failback plan is approved.