Having a backup does not tell you whether a suitable version will exist when you discover a problem weeks later. Retention tiers combine frequent recent copies with less frequent long-term recovery points.
Start with loss scenarios
Map deletion, corruption, ransomware, an incorrect deployment and delayed discovery. Each scenario requires a different depth of history.
Define RPO and RTO for each system
How much data can you afford to lose, and how quickly must service return? Do not apply one expensive level everywhere.
Design daily, weekly and long-term tiers
Keep recent restore points closer together and older snapshots further apart. Document exactly when rotation and deletion happen.
Protect against the same attacker
Use isolation, immutability and separate credentials. An online backup with shared access can be deleted along with production.
Test an actual restoration
Confirm integrity, keys, dependencies and restoration time. A successful backup job does not prove that the service can be recovered.
This framework is an original editorial methodology developed by DigitalNow.