Backups and Point-in-Time Recovery
~8 min read
Automated daily snapshots plus continuous transaction log backup enable restoring to any specific second.
RDS automated backups combine a daily full snapshot with continuous backup of transaction logs throughout the day, enabling point-in-time recovery to any specific second within your configured retention window (1 to 35 days) — not just to the moment of the last daily snapshot. This is critical for recovering from application-level mistakes (e.g. a bad migration script that corrupted data at 2:47pm) where you need to restore to the exact moment before the mistake, not just 'sometime yesterday.'
A point-in-time restore always creates a NEW RDS instance rather than overwriting the existing one — this is a deliberate safety design, letting you inspect and validate the restored data before deciding to cut over application traffic to it, rather than risking a destructive in-place restore.
Manual snapshots, unlike automated backups, persist indefinitely (not subject to the retention window) and are useful before risky operations like a major version upgrade or schema migration, giving you an explicit rollback point independent of the automated backup schedule.
In the AWS Console
- 1
RDS → Databases → select instance → Actions → Restore to point in time
Choose the exact date and time to restore to, and a new DB instance identifier for the resulting restored instance.
The restored instance is entirely new and separate — you'll need to update application connection strings to switch to it after validating the data.
💬 Deep Dive with AI
Key points
- •Automated backups = daily snapshot + continuous transaction logs, enabling restore to any second within the retention window
- •Retention window is 1-35 days, configurable based on recovery point objective
- •A restore always creates a new instance — never overwrites the existing one in place
- •Manual snapshots persist indefinitely, useful as an explicit checkpoint before risky operations