Backups and snapshots
Snapshot schedules, retention, restores and restore drills.
How snapshots work
A snapshot captures the entire boot disk at a point in time. Snapshots are incremental after the first one, stored on separate storage, and replicated within the region. Taking a snapshot does not require downtime.
Schedules and retention
Enable automatic snapshots per server in the console. Choose a daily window and a retention period of 7, 30 or 90 days. Older snapshots are deleted automatically after the retention window.
| Plan | Included snapshots | Included storage |
|---|---|---|
| Starter | Add-on ($1/month) | 20% of disk |
| Growth | Daily | 20% of disk |
| Scale | Daily | 20% of disk |
Storage beyond the included amount is $0.02/GB-month. On-demand snapshots are unlimited but count toward storage.
Restore a snapshot
- Open the server and select Snapshots.
- Choose a snapshot and click Restore.
- Restore in place, or create a new server from the snapshot to inspect it first.
When restoring in place, we take a safety snapshot of the current disk first, so a mistaken restore can be undone. The server reboots during the operation.
br snapshots create api-prod --label "before-upgrade"
br snapshots list api-prod
br snapshots restore api-prod --snapshot snap_8fk2 --target new
Restore drills
Every month we restore a sample of production snapshots into an isolated network and verify that the machines boot and serve traffic. If a drill fails, we treat it as an incident and fix the underlying cause.
Deleting servers and snapshots
Destroying a server keeps its snapshots until their retention date. Deleting a snapshot is permanent. If a snapshot is locked for compliance, it cannot be deleted before its lock expires.
