Why it matters
Most file systems hold a lot of data that is rarely read, and lifecycle management is off unless someone turns it on.How nOps finds it
nOps reads the last 30 days of EFS billing. A file system is flagged when its Standard (or One Zone) storage was billed on at least 28 of those days, it averaged at least 100 GB over the last 7 days, and it had no Infrequent Access or Archive storage or access in 30 days. File systems whose provisioned throughput costs more than their storage are skipped. The cold share is the median share of data in Infrequent Access or Archive across your file systems that already use lifecycle management, and at least 3 of them (each 100 GB or more) are needed. The saving is the Standard storage cost times that share times the price difference between Infrequent Access and Standard, and must be at least $5 a month.Before you start
Permission to update lifecycle configuration (elasticfilesystem:PutLifecycleConfiguration).
How to implement
- In the EFS console, open the file system and choose Edit under General.
- Set Transition into IA to 30 days since last access and Transition into Standard to On first access.
- Or use the CLI:
aws efs put-lifecycle-configuration --file-system-id ID --lifecycle-policies TransitionToIA=AFTER_30_DAYS TransitionToPrimaryStorageClass=AFTER_1_ACCESS. - Update Terraform or CloudFormation with the same policy.
Risks and considerations
- Moving data into Infrequent Access has a one-time charge per GB, and reads from Infrequent Access are charged per GB.
- Files in Infrequent Access have slightly higher latency on first read.
- The estimate assumes this file system has a similar share of cold data to your others.
- To roll back, turn lifecycle management off and set files to move back on first access.