AWS cost optimization

Welcome to Cloud Cost Clinic: Safe AWS Cost Cleanup

Cloud Cost Clinic helps small AWS teams find possible waste, verify ownership and dependencies, and reduce AWS spend without risky cleanup guesses.

Published 2026-05-18 Updated 2026-05-18 By Cloud Cost Clinic

Quick takeaway

Cloud Cost Clinic is a practical AWS cost optimization project for small teams that need to understand the bill, find possible waste, and clean up safely. The core idea is simple: a high bill is a signal, not a deletion order.

The goal is not to delete more AWS resources. The goal is to make every cleanup decision easier to explain, safer to approve, and less likely to break production.

This is where I put on the serious billing cardigan: cost cleanup is still a production change when production can feel it.

Cloud Cost Clinic safe cleanup reminder about checking ownership, dependencies, and rollback before deleting AWS resources

Why this exists

AWS cost advice often gets noisy. One article says to right-size everything. Another says to delete unused resources. A dashboard points at a spike. A budget alert lands in an inbox. None of those things automatically answer the important question: what is safe to change?

Small teams usually do not have a dedicated FinOps team watching every line item. The person reviewing the bill may also be shipping features, running deployments, debugging incidents, and trying to remember why a staging database still exists.

That staging database may have a reason. It may also have a very confident bill. The review exists because both can be true.

Cloud Cost Clinic is built for that reality. It turns AWS cost review into a practical workflow:

  • Start with the bill.
  • Narrow the signal by service, usage type, account, Region, and tag.
  • Mark findings as possible waste, not confirmed waste.
  • Find the owner and operational context.
  • Choose a safe next action.
  • Document the decision so the same mystery does not return next month.

What small teams usually miss

Most AWS waste in small accounts does not come from one dramatic mistake. It comes from ordinary infrastructure that outlives its original purpose.

Common patterns include:

  • Unattached EBS volumes from old migrations, tests, or incidents.
  • Old EBS snapshots kept long after restore expectations changed.
  • CloudWatch log groups with no retention policy.
  • NAT Gateway and data transfer costs caused by traffic paths nobody reviews.
  • RDS instances sized for a past peak or a future that did not arrive.
  • Idle load balancers, public IPv4 addresses, and temporary environments.
  • S3 buckets without lifecycle rules or clear restore expectations.
  • Untagged spend where nobody can tell who owns the resource.

Any of these can be waste. None of them are automatically safe to remove.

The Cloud Cost Clinic rule

Possible waste is not the same thing as safe cleanup.

That rule is the center of the site. A resource can look idle and still be part of a rollback plan. A snapshot can look old and still support restore expectations. A database can look quiet and still carry a reporting workload, batch job, or seasonal pattern.

Cost Explorer can tell you where to look. It cannot tell you what will break. A service console can show whether something is attached. It cannot tell you whether someone expects it to be there during recovery.

That is why every Cloud Cost Clinic guide follows the same pattern:

  • What is the cost signal?
  • Why can it create AWS spend?
  • How do you find review candidates?
  • What should you verify before changing anything?
  • What safer action can you take first?
  • What should you avoid?

How to start a safer AWS cost review

Start with a short review queue instead of a cleanup sprint. The first version can be a spreadsheet, ticket list, or shared document. It only needs enough structure to slow down bad guesses.

Use these fields:

  • Finding: what looks unusual, expensive, idle, old, or ownerless?
  • Service: which AWS service or billing group raised the signal?
  • Account and Region: where does the cost appear?
  • Evidence: what Cost Explorer view, service metric, tag, or trend supports the finding?
  • Owner: who can explain the workload or approve a change?
  • Risk: low, medium, or high based on production impact and reversibility.
  • Next action: review, tag, retain, expire, resize, stop, delete, reroute, or accept.
  • Rollback path: what happens if the change causes a problem?

This turns vague billing anxiety into a decision list.

Safe first actions

The safest first action is often not deletion. It is making the next decision clearer.

Good first actions include:

  • Add missing owner and environment tags.
  • Create or tune AWS Budgets alerts.
  • Save useful Cost Explorer reports for monthly review.
  • Set agreed retention for non-production log groups.
  • Add expiration dates to temporary environments.
  • Mark accepted costs so they are not re-litigated every month.
  • Review low-risk non-production resources before touching production.

Deletion, resizing, retention changes, and route changes can be valid next steps. They just need ownership, evidence, dependency checks, timing, and rollback.

What this site will cover

The first Cloud Cost Clinic series focuses on common AWS cost leaks small teams miss:

  • NAT Gateway and data transfer surprises.
  • Unattached EBS volumes.
  • Old EBS snapshots.
  • Idle Elastic IPs and public IPv4 spend.
  • CloudWatch log retention.
  • Overprovisioned RDS.
  • Idle load balancers.
  • S3 storage classes and lifecycle rules.
  • Cross-AZ and data transfer paths.
  • Missing budgets, tags, and owners.

The guides are educational. They use public AWS documentation and demo-style examples, not private billing exports, employer data, client examples, account IDs, ARNs, or internal screenshots.

What not to do

Do not treat "unused" as "safe to delete." It is a reason to investigate.

Do not chase the biggest possible savings estimate before checking the blast radius. A smaller reversible change can be more useful than a dramatic cleanup.

Do not let a dashboard replace ownership. Another chart will not answer who understands the workload.

Do not publish exact AWS pricing claims without checking current AWS pricing pages. Pricing varies by Region, service configuration, usage, date, commitments, and discount programs.

Do not turn a monthly bill review into a production change with no change window.

Checklist

  • Start with Cost Explorer or the bill, not a random resource list.
  • Group spend by service, usage type, linked account, Region, and tag.
  • Separate spikes from recurring baseline spend.
  • Mark review candidates as possible waste.
  • Find the owner before recommending a change.
  • Check dependencies, backup needs, restore expectations, and usage signals.
  • Prefer reversible or low-risk changes first.
  • Schedule production changes with a rollback path.
  • Document accepted costs so the same item has context next month.

FAQ

What is Cloud Cost Clinic?

Cloud Cost Clinic is a pseudonymous educational project about practical AWS cost optimization for small teams. It focuses on finding possible waste, verifying context, and making safer cleanup decisions.

Is this affiliated with AWS?

No. Cloud Cost Clinic is not affiliated with AWS, any employer, any client, or any vendor. AWS service behavior and billing claims should be checked against official AWS documentation and pricing pages.

Should small teams delete unused AWS resources?

Resources that look unused are review candidates, not automatic deletion targets. Check owner, workload, environment, dependencies, backups, restore needs, and rollback before removing anything.

What is the best first AWS cost habit?

Create a small monthly review queue. Start with Cost Explorer, group top spend by usage type and ownership, then record findings with evidence, risk, owner, and next action.

Sources

Reader question

If you reviewed your AWS bill once a month, what would help most first: a checklist, a spreadsheet, a report template, or a small read-only scanner?