The AWS bill that grew while nobody was looking

Finance sends the screenshot on a Monday. The line item is up. Nobody shipped a new product. The staging account did what staging accounts do: it kept the lights on.

A three-person platform team left a few m5s running after a load test. Three gp3 volumes never got attached again. An Aurora instance sat there "in case we need the dump." Elastic IPs nobody routes to. None of this is dramatic. All of it is a monthly charge.

That is the usual story. Not a crypto miner. Not a misconfigured Lambda. A test that ended, a person who changed teams, a region nobody opens in the console.

What Cost Explorer will not tell you

Cost Explorer is good at "EC2 in eu-west-1 went up." It is weak at "this volume has been available for 40 days and still bills." Trusted Advisor and Compute Optimizer help on a subset of services, if you remember to open them, if the support plan unlocks the check, if someone owns the ticket.

The waste that hurts most is boring:

• EC2 that idles all week
• RDS stopped for more than a week (you still pay storage)
• NAT gateways with almost no traffic
• SageMaker endpoints left after a demo
• Snapshots and AMIs kept "just in case"

What we actually do

unusd.cloud is the audit I used to run by hand for clients. You deploy a read-only IAM role (CloudFormation, Terraform, or StackSets). We scan 30+ services in every enabled region with 51+ fixed rules. Nothing in the account changes unless you change it.

You get a weekly email you can forward. Slack or Teams for the scan. The full report and Navi live in the app. Leadership can get a short org digest by email or Telegram. Setup is about five minutes.

We do not invent a "you will save 20 percent" number. The first scan lists resources, regions, and a monthly estimate. You decide what to stop, delete, or whitelist.

If you want the same pass on one account, start a free scan. No credit card. For how this sits next to Cost Explorer, read the comparison.