# Enable Cloud Backup

## Steps

1. Atlas UI: Backup, pick the cluster, enable Cloud Backup. (Available on M10+ dedicated clusters; Free and Flex have their own limited backup story.)
2. Set the snapshot schedule: hourly snapshots with a day or two of retention for tight RPO, plus daily snapshots kept a week, plus weekly kept a month. Match retention to how far back a "we need that data" request can reach.
3. Turn on point-in-time restore (continuous oplog). This is what lets you restore to any second inside the oplog window instead of only to snapshot times.
4. If you operate in multiple regions, configure snapshot copy to the regions you would rebuild in.

## Rules

- Disabling continuous backup deletes the backup history. Do not toggle it casually.
- Snapshot storage and cross-region copies cost money; size retention deliberately.
- The oplog window bounds point-in-time restores. Know your window and alert if it shrinks.

## Verify

Trigger an on-demand snapshot and watch it complete. Then do a trial restore to a scratch cluster (never over production) and confirm the data is there. A backup you have never restored is a hope, not a backup.