Compute is frequently 40%+ of the bill

AWS EC2 right-sizing and idle instance cleanup

Compute is often the largest slice of an AWS bill and the easiest to over-provision. Right-sizing, current-generation and Graviton instances, and killing idle dev boxes routinely cut EC2 spend by a third.

Why the waste happens

  • Instances are sized for a launch-day peak or a worst case that never recurs, then never scaled down.
  • Dev, test, and staging instances run 24/7 when they are only used during working hours.
  • Workloads stay on previous-generation families (m4, c4, r4) instead of cheaper, faster current-gen (m7, c7, r7).
  • Instances run on Intel when Graviton (arm64) would be about 20 percent cheaper for the same work.
  • The steady baseline runs entirely on-demand with no Savings Plan or Reserved Instance coverage.

How to spot it

  • Instances with consistently low CPU and network utilization over several weeks.
  • Previous-generation instance types still in use.
  • Non-Graviton instances running workloads that have arm64 builds available.
  • Non-production instances running overnight and on weekends.

The fix

  • Right-size to the smallest instance that holds your real peak, then buy a Savings Plan for the steady baseline.
  • Schedule stop/start (or auto-scaling to zero) for dev and test instances outside working hours.
  • Migrate to current-generation and Graviton families for a same-or-better-performance discount.
  • Use Spot Instances for fault-tolerant and batch workloads.

Find every EC2 issue in one scan

Cloud Cost Analyzer checks every instance against rules for idle and oversized compute, previous-generation families, Graviton migration candidates, stopped instances still holding EBS, burstable misuse, and Savings Plan or Reserved Instance coverage, each with a dollar estimate.

$ cca scan --provider aws

Read-only, runs in about 5 minutes. Free tier, no credit card.

EC2 cost FAQ

How do I reduce my AWS EC2 costs?+

Right-size over-provisioned instances, migrate to current-generation and Graviton families, schedule non-production instances to stop outside working hours, use Spot for fault-tolerant work, and cover the steady baseline with a Savings Plan or Reserved Instances.

Is migrating EC2 to Graviton difficult?+

If your workload has an arm64 build (most modern runtimes and containers do), it is usually a straightforward instance-type change. Graviton is typically about 20 percent cheaper for equivalent performance. Test in a non-production environment first.

Should development EC2 instances run 24/7?+

Usually not. Dev and test instances are often used only during working hours, so scheduling them to stop nights and weekends can cut their cost by 60 to 70 percent with no impact on developers.

See what you are overpaying for

Point a read-only CLI at your account and get a ranked fix list with dollar amounts. Free tier, no credit card.