Skip to main content
S3 Standard-IA stores data for less than S3 Standard, but charges for every GB retrieved and for objects deleted before 30 days. When data in Standard-IA is read often or deleted soon after it moves, those fees can cost more than the discount. nOps finds buckets where keeping the data in S3 Standard would cost less.

Why it matters

Lifecycle rules that move everything to Standard-IA after a fixed number of days are common, and they backfire on data that is still in use.

How nOps finds it

nOps reads the last 30 days of S3 billing. A bucket is flagged when it has at least $5 of Standard-IA storage, and its Standard-IA storage plus retrieval and early-delete fees exceed what the same data would cost in S3 Standard by at least $25 a month. The Standard cost uses the public price ratio between Standard and Standard-IA in the region. The saving is that difference.

Before you start

Permission to update lifecycle rules (s3:GetLifecycleConfiguration, s3:PutLifecycleConfiguration).

How to implement

  1. Find the rule that transitions data to STANDARD_IA: aws s3api get-bucket-lifecycle-configuration --bucket BUCKET.
  2. Remove that transition, or change it to INTELLIGENT_TIERING, which has no retrieval fees.
  3. Write back the full edited configuration with put-bucket-lifecycle-configuration, since it replaces every rule.
  4. Optionally copy frequently read objects back to Standard.
  5. Update Terraform or CloudFormation with the same change.

Risks and considerations

  • Objects already in Standard-IA stay there unless you copy them back. Copying them early can itself trigger early-delete fees.
  • To roll back, restore the original lifecycle rule.