Skip to content

S3

Amazon S3 is cheap per gigabyte until a data lake, log archive, or query-result dump sits in Standard for years. Size is not the same thing as waste: those buckets are usually still needed.

Implementation Effort: Medium - Estimated time: 2-4 hours. Add lifecycle or Intelligent-Tiering. Do not delete the bucket because it is large.

What We Detect

  • Largest buckets -- Your top buckets by storage size and current monthly storage cost, as a spend map
  • Storage class mix -- How much data is already in Glacier, Intelligent-Tiering, or Deep Archive versus Standard
  • Archive and expiry opportunities -- Buckets still in Standard with no lifecycle or Intelligent-Tiering, with a conservative savings estimate. Logging and Athena result buckets are candidates to expire objects, not to delete the bucket

Why It Matters

S3 Standard is billed whether or not anyone reads the objects. Moving cold data to archive, or expiring query results after a retention window, can cut that bill without throwing the bucket away. Pricing the whole bucket as if you would delete it overstates savings and is not a valid recommendation.

Recommendations

  1. Keep the bucket -- a large logging, backup, or data-lake prefix is almost never unused
  2. Add Intelligent-Tiering or a lifecycle rule so Standard objects can archive as they cool
  3. Expire objects on logging and Athena query-result buckets after the retention you actually need

Keep on chasing 🧡