Backup types and retention
The Backups page reports the protection currently available for an App. How backups and restores work explains how these layers relate.
| Protection | What it covers | Customer-visible retention |
|---|---|---|
| Database backups | Full database recovery points | The latest 14 backups |
| Point-in-time recovery | Database changes between recovery points | The restorable time window shown in the portal |
| App-file backups | Persistent App workspace files | Expiration shown for each recovery point |
| Media versions | Overwritten or deleted media objects | 30 days for prior versions |
| Cross-region protection | Eligible database, App-file, and media copies | Health and last-check status shown in the portal |
| Restore safety backup | Database state immediately before a restore | Shown as the restore’s Undo point |
Count-based and time-based retention
Section titled “Count-based and time-based retention”Database backup retention is count-based. The oldest backup expires when a new backup exceeds the retained count, so individual database backups don’t display expiration dates.
Media retention is time-based. The current object remains available while the App exists; overwritten or deleted prior versions expire after 30 days.
Point-in-time recovery
Section titled “Point-in-time recovery”The portal displays the earliest and latest eligible restore times. Select a timestamp inside that range. The range can move forward as older recovery data expires.
Status freshness
Section titled “Status freshness”Backup mechanisms update independently. The page uses the oldest available status timestamp when summarizing freshness. A stale check means the displayed state needs confirmation; it doesn’t by itself prove that the underlying backup has disappeared.
Retention shown in the portal is authoritative for the App. Contractual or legal retention outside the customer restore surface doesn’t create an additional self-service recovery point.