Inventory before moving
List services, scheduled jobs, persistent storage, ports, and external dependencies. Record where secrets are managed without copying them into the runbook. Assign a person to the cutover decision.
Prove the backup can restore
Create a consistent backup and test restoring it in an isolated environment. Record the recovery point and estimated recovery time. A backup file existing is not evidence that it can restore the service.
Validate the destination
Check application health, database connectivity, background jobs, and storage permissions. Verify the real hostname and HTTPS certificate against the new destination using a local hostname override. A direct-IP response alone is insufficient.
Plan the write boundary
Decide how writes will stop or replicate during cutover. Avoid two independent systems accepting writes unless you have a conflict strategy. Record a final sync step and compare counts and important records afterward.
Cut over with a rollback gate
Confirm the provider’s completion notice and your own health checks before changing DNS. Monitor the public hostname, application errors, and background jobs. Keep the old environment available for the agreed rollback window, taking care not to resume conflicting writes.
Further reading
Put it into practice with our free browser tools.
Open the toolbox →