Disaster Recovery

Have a Plan B for disaster recovery

During a disaster recovery you want to be fast to respect your RTO .

It is possible, for example, to restore the operation of a service using snapshots (both volumes and instances).

Restoring snapshots may fail for unforeseen reasons (e.g., you need to restore the service on a different cloud provider), so you should make sure that you can restore from scratch (ability to recreate the service, use database dump, etc.).

This could take longer and may lead to non-compliance with your RTO, but it’s still preferable to be able to restore with a slight delay (resulting in payment of a penalty or customers being given discounts) rather than not being able to recover the system at all.