3.2.1.8. Cost-Optimized Backup Retention (AWS Backup, Data Lifecycle Manager)
3.2.1.8. Cost-Optimized Backup Retention (AWS Backup, Data Lifecycle Manager)
💡 First Principle: Backup spend is the number of recovery points times their size times the storage tier times how long they are kept. Retention is therefore a cost decision that must be derived from the recovery objectives and the compliance rule, with each tier of retention kept in the cheapest place that still meets its restore time.
Scenario: A company keeps daily backups of "EFS", "DynamoDB" and "EBS" for seven years in warm storage, and copies every one of them to a second Region, because the compliance team asked for "seven years of backups." The actual requirements are 35 days of daily restore points, monthly points for seven years, and a weekly cross-Region copy for disaster recovery.
"AWS Backup" plan design:
- Tier the rules: Use separate backup rules for each retention need (for example daily kept 35 days, monthly kept 7 years). When rule windows overlap, AWS Backup keeps the backup from the rule with the longer retention rather than taking two.
- Cold storage: A rule's lifecycle can move recovery points from warm to cold storage. Cold storage is supported for
"EFS","DynamoDB"(with AWS Backup advanced features),"EBS"(through EBS Snapshots Archive),"Timestream", VMware VMs, SAP HANA on"EC2"and"CloudFormation"stacks, not for"RDS"or"Aurora", whose backups stay warm. A backup must stay in cold storage at least 90 days, so total retention must be at least 90 days more than the transition setting. Moving daily backups with 35-day retention to cold is therefore not possible; moving monthly backups after 30 days is the classic saving. AWS recommends at least 8 days in warm storage for incremental resources so a warm full backup remains. - Copies cost twice: A cross-Region or cross-account copy action stores a second set of recovery points and pays transfer. Copy only the rule that the DR objective needs (for example weekly), and give the copy its own shorter expiry.
- Organization-wide:
"AWS Backup"policies in"AWS Organizations"push the same plans to every account, and"Vault Lock"makes retention immutable for compliance, so cost-optimized retention does not depend on each team configuring it correctly.
"EBS" snapshots and "Amazon Data Lifecycle Manager":
- EBS Snapshots Archive: Up to 75% lower storage cost for snapshots kept 90 days or more and rarely restored. Archiving converts an incremental snapshot into a full snapshot, restores take up to 72 hours, and deleting before 90 days still bills the minimum. It saves money for monthly, quarterly, yearly or end-of-project snapshots; archiving a daily chain, where each incremental snapshot holds little changed data, can cost more than leaving it in the standard tier.
"Data Lifecycle Manager"(DLM): A no-charge, tag-targeted scheduler for"EBS"snapshots and EBS-backed AMIs that keeps them by count or age, copies them cross-Region with their own retention, and can archive them (an archive rule needs a schedule of 28 days or longer and at least 90 days in the archive tier). Choose DLM for EBS-only, tag-driven retention inside an account; choose"AWS Backup"for many services, central vaults, cross-account copies, organization backup policies and Vault Lock.- Clean up the long tail: Manual
"RDS"snapshots and manually created"EBS"snapshots never expire on their own, snapshots of deleted volumes keep billing, and a long"Recycle Bin"retention keeps deleted snapshots billing too. Find them before buying anything.
⚠️ Exam Trap: Choosing "transition all backups to cold storage after 1 day" as the cheapest option. Cold storage requires 90 days there, does not apply to "RDS"/"Aurora", and for "EBS" stores full rather than incremental copies, so it only saves money for long-retention, rarely restored recovery points.