A retention policy determines how long backup snapshots remain available for recovery. In Atlas Cloud Backup, it works alongside the snapshot schedule: the schedule determines when copies are taken, while retention determines how long they are kept. Longer retention can support recovery and compliance needs. section 21.1.1 section 21.5
For a dedicated M10-or-higher cluster, inspect the current schedule with `atlas backups schedule describe <cluster-name>`. Read each row’s frequency and retention together; this command describes the settings rather than changing them. section 21.1.1
CLUSTER NAME AUTO EXPORT ENABLED NEXT SNAPSHOT MongoDB-in-Action-M10 false 2024-06-22 14:49:20 +0000 UTC ID Frequency Interval Frequency Type Retention Value Retention Unit 6676900fe259a12fa8ccfc73 6 hourly 7 days 6676900fe259a12fa8ccfc74 1 daily 7 days 6676900fe259a12fa8ccfc75 6 weekly 4 weeks 6676900fe259a12fa8ccfc76 40 monthly 12 months 6676900fe259a12fa8ccfc77 12 yearly 1 years
This example schedule lists hourly, daily, weekly, monthly, and yearly snapshots with separate retention values and units. For example, its hourly snapshots are retained for seven days. section 21.1.1
Choose the backup approach before relying on a retention window:
- Dedicated M10-or-higher clusters support scheduled snapshots with configurable retention and on-demand snapshots. Continuous cloud backup offers point-in-time recovery within a configured retention window, with possible additional cost. section 21.5
- Flex Cloud Backup is limited to automatic daily snapshots; M0 clusters cannot use Atlas Cloud Backup. Manual `mongodump` backups are an alternative for M0, but Atlas does not support its `--oplog` option for point-in-time captures. section 21.1.1 section 21.5
- Do not treat Atlas Online Archive as a backup retention policy. Its rules address access patterns and retention requirements for lower-cost storage, not recovery after data loss or corruption. section 17.0
