What is your AWS storage actually costing you?

Versioning, incomplete multipart uploads, and snapshot-like accumulation are the usual causes. None of them show up in a normal object count, and none of them stop billing until a lifecycle rule explicitly expires them. Check for a bucket with versioning on and no noncurrent-version expiration rule first; it's the single most common source of silent S3 growth.

For unpredictable access patterns, usually yes. Its three automatic tiers are priced identically to the manual classes they mirror, so the storage cost lands in much the same place. But it charges no retrieval fee on any of them and carries no early-delete penalty, which means guessing wrong about an access pattern costs you nothing beyond the monitoring fee. The trade is that per-object monitoring fee, plus no benefit at all for objects under 128 KB, which never get auto-tiered. If you already know an access pattern for certain (backups nobody will touch again for a year, say), a manual lifecycle rule into Glacier Flexible Retrieval or an opt-in Archive tier can still beat it on pure storage cost.

The storage rate is identical. The difference is who decides when data moves: with Standard-IA you write the lifecycle rule and take the early-delete risk if you're wrong about access patterns; with Intelligent-Tiering, AWS moves objects automatically after 30 days of no access and moves them back the moment they're read again, with no early-delete penalty and no retrieval fee either way.

EBS snapshots are incremental. Each one only stores the blocks that changed since the last snapshot in its chain, even though the console shows every snapshot's full logical size. Deleting one snapshot only frees the blocks no remaining snapshot in that chain still depends on, which is often far less than its listed size implies.

It's an online, no-downtime operation for an existing volume, and at equal provisioned size it's a 20% saving with a higher default performance baseline (3,000 IOPS / 125 MiB/s free) than gp2 offered at the same size. The one thing worth checking first is whether anything was relying on gp2's size-scaled IOPS ceiling. gp3 decouples performance from size, so a very large gp2 volume that had accumulated a high implicit IOPS ceiling may need extra provisioned IOPS on gp3 to match it.

You can restore or delete it any time. Nothing blocks the action, but you'll be charged the early-delete penalty: the storage rate for whatever days remain in the class's minimum (90 for Glacier Instant Retrieval and Glacier Flexible Retrieval, 30 for Standard-IA and One Zone-IA), on top of anything you already paid.

No, deliberately. Every other figure on this page came from a first-party pull of AWS's Price List Bulk API, and per-request pricing wasn't part of that pull. Rather than estimate it, the field was left out entirely. Everything shown here is a storage, retrieval, or monitoring cost only.

No. CostMon doesn't rewrite a lifecycle policy, delete a volume, or migrate a gp2 fleet to gp3 for you. What it does is pull your real AWS cost and usage data on a schedule, normalize it alongside every other provider you run, and flag a climbing storage line before it turns into a surprise on the invoice. The fixes above are still yours to apply.