Data is backed up on a regular schedule to secure, separate storage, so recent data can be restored if something goes wrong.
Backups are checked periodically to confirm they can actually be restored -- a backup that cannot restore is not a backup.
We aim to restore service within [target time, RTO] and to limit data loss to [target window, RPO]. [Set real targets for your setup.]
If primary systems fail, we restore from backups onto healthy infrastructure and communicate status to affected users.