Amazon S3 Costs: Storage Classes, Lifecycle and Access

Why a schedule does nothing here

S3 is object storage with no running compute to switch off. AWS lists its cost components as storage, requests and data retrieval, data transfer and transfer acceleration, management and insights features, replication, and transform and query features. Every one of those is driven by how much you keep, in which class, and how it is read. So the work on S3 is about the data, not the clock.

Where S3 spend actually hides

• Objects in the wrong class. Data that nobody reads sitting in S3 Standard.
• No lifecycle rules. Logs and exports that were meant to age out but never do.
• Old versions. Versioned buckets keep every overwritten copy until something expires it.
• Abandoned multipart uploads. Parts of uploads that never completed still occupy storage.
• Features left on. Transfer Acceleration enabled on a bucket with little cross-region traffic.

What ZopNight reads and recommends

ZopNight discovers buckets on each connected account and enriches them with bucket size and object count from CloudWatch, so a bucket in the list carries its real footprint. Cost is shown per bucket, as rack rate or as billed cost once billing data is connected.

The S3 rules map onto the problems above: S3 Storage Tier Migration, S3 Intelligent-Tiering Not Enabled, S3 Bucket Lifecycle Policy Missing, S3 Noncurrent Version Accumulation, S3 Incomplete Multipart Uploads Accumulating, S3 Glacier Objects, Consider Deep Archive, S3 Transfer Acceleration Enabled but Low Cross-Region Traffic, and Large S3 Bucket With External Access, Consider Requester Pays. On the access side, S3 Bucket With Public Access Enabled, S3 Bucket Not Requiring HTTPS and S3 Bucket Access Logging Not Enabled.

Each finding carries the evidence that triggered it and a monthly saving where one can be computed honestly. When a figure cannot be proven the rule abstains rather than invent one.

Where S3 meets the rest of your bill

Buckets are rarely the only storage line. EBS snapshots are stored in S3 and billed by the space they consume, and NAT gateway traffic to S3 is its own rule family on the networking side. The EBS page covers snapshots.

Details per bucket type live on the S3 fact sheet. For the method behind class choices, see storage tiering, and for the broader cleanup pattern, orphan cleanup.