# Atlas Cloud Backup (snapshots) > Source-anchored research for Atlas Cloud Backup (snapshots). Run status: COMPLETED. Parent: [MongoDB Backup and Restore](https://llms-explorer.com/tree/mongodb-backup-and-restore/) · 3 facets · 15 facts · page: https://llms-explorer.com/tree/atlas-cloud-backup-snapshots/ ## How it works - For a single-region Atlas Core cluster, Atlas attempts incremental snapshots when possible; if the existing snapshot storage volume becomes invalid, Atlas creates a new volume and takes a full-copy snapshot before continuing with incremental snapshots. (low confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/backup/cloud-backup/dedicated-cluster-backup.md) - A Backup Compliance Policy protects backups from user deletion and prevents users from lowering protected retention; snapshots in an optional extra-retention period become deletable again, then Atlas deletes them when that period ends. (medium confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/backup/cloud-backup/configure-backup-policy.md) - When Cloud Backup is disabled or a cluster is terminated, Atlas lets the operator retain or delete existing snapshots; Prisma's 2026-09-30 guide says cluster deletion removes snapshots unless retention is selected or a Backup Compliance Policy preserves them. (medium confidence) (verified as of 2026-09-30) — [source](https://www.mongodb.com/docs/atlas/backup/cloud-backup/configure-backup-policy.md) ## How-to and procedures - To apply a changed retention period to existing snapshots, an administrator must select the option to update existing snapshot retention before saving the policy; the change only applies to snapshots from the updated policy items whose retention was not changed individually through the API. (low confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/backup/cloud-backup/configure-backup-policy.md) ## Facts and statements - For Atlas Core, recurring Cloud Backup snapshots follow policy items that define the UTC snapshot time, frequency interval, and retention duration; Atlas also supports taking snapshots outside the schedule. (high confidence) (verified as of 2026-09-30) — [source](https://www.mongodb.com/docs/atlas/backup/cloud-backup/configure-backup-policy.md) - Atlas Cloud Backup uses localized snapshots through each cluster's cloud provider; Atlas says incrementality is provider-dependent and calculates monthly storage from provider-reported snapshot metrics and region. AWS documents EBS snapshots as full initially and incremental afterward, retaining changed and new blocks. Atlas also lists restore-transfer charges and extra VM, restore-volume, and transfer charges for manual HTTPS downloads. (medium confidence) (verified as of 2026-10-20) — [source](https://www.mongodb.com/docs/atlas/billing/cluster-configuration-costs/) - Atlas Core can restore scheduled or on-demand snapshots to selected databases or collections. A cluster-level restore deletes the target's existing data and makes the target unavailable during restoration. For sharded restores, source and target must have the same shard count and config server type. (high confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/backup/cloud-backup/restore-from-snapshot.md) - If a scheduled snapshot fails, Atlas retries and may provide a fallback snapshot. Atlas warns that fallback snapshots may be inconsistent across the cluster and should be a last-resort restore path; Atlas does not allow fallback snapshots to be exported to object storage. (high confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/backup/cloud-backup/restore-from-snapshot.md) - Atlas Cloud Backup snapshots can be exported manually or automatically to AWS S3 or Google Cloud Storage. Automatic exports include only snapshots whose frequency matches the configured backup policy item, and the export documentation specifies M10 or higher clusters. (high confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/backup/cloud-backup/export/aws-s3.md) - Backup availability and control differ by Atlas tier: Free clusters have no Atlas backups, Flex clusters receive automatic daily snapshots with a fixed schedule, and Dedicated M10+ clusters use configurable Cloud Backup policies. (high confidence) (verified as of 2026-10-15) — [source](https://www.mongodb.com/docs/atlas/backup-restore-cluster.md) - Atlas Cloud Backup uses cloud-provider snapshot storage; for single-region Atlas Core clusters, Atlas stores snapshots in the same cloud region as the cluster. (medium confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/backup/cloud-backup/dedicated-cluster-backup.md) - Atlas Core Cloud Backup snapshots are immutable by default; Atlas Infinite backups are not immutable. (low confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/backup-restore-cluster.md) - Atlas Core supports restoring selected databases or collections from a scheduled or on-demand snapshot. (high confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/backup/cloud-backup/restore-from-snapshot.md) - A cluster-level restore deletes existing data on the target and makes that target unavailable during the restore; Atlas requires the target to receive no client requests, so restoring to a new cluster and switching application traffic afterward is an available cutover approach. (high confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/backup/cloud-backup/restore-from-snapshot.md) - Atlas restores a snapshot to a target cluster running the same MongoDB major version with an equal or higher minor version, or the next higher major version; if the snapshot has a pinned FCV, the target major version must match that FCV. (medium confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/backup-restore-cluster.md)