AWS cost optimization: 12 checks to cut waste from your bill

Almost every AWS bill carries avoidable spend. It rarely comes from one big mistake. It builds up from many small defaults that nobody revisits. Here are the twelve checks our FinOps engineers run first.

Quick wins: the first week

1. Delete idle and orphaned resources. Unattached EBS volumes, old snapshots, unused Elastic IPs, idle load balancers and stopped instances still holding storage all cost money every month.
2. Rightsize compute. Compare CPU and memory utilization in CloudWatch or AWS Compute Optimizer with the instance size you pay for. Instances that rarely pass 20% CPU are candidates to step down a size.
3. Move EBS volumes from gp2 to gp3. gp3 costs less per GB than gp2 and lets you set IOPS and throughput independently. The change usually needs no downtime.
4. Schedule non-production environments. Development and test environments that only run during working hours can be stopped nights and weekends, cutting their compute cost by more than half.

Structural savings: the first month

1. Commit to your baseline. Savings Plans or Reserved Instances for steady workloads typically cost much less than On-Demand. Cover the stable baseline, not your peaks.
2. Consider Graviton. AWS Graviton (ARM) instances generally offer better price-performance for many Linux workloads, including containers, web servers and databases on RDS.
3. Tier your storage. S3 Intelligent-Tiering or lifecycle rules move rarely used objects to cheaper storage classes automatically.
4. Watch data transfer and NAT gateways. Cross-AZ traffic, NAT gateway processing and internet egress are often the most surprising lines on the bill. VPC endpoints for S3 and DynamoDB can remove a large share of NAT costs.
5. Set log retention. CloudWatch Logs kept “forever” by default grow quietly. Set retention periods and archive what you must keep.

Make savings stick

1. Tag everything. Cost allocation tags by team, product and environment turn one large bill into numbers people can own.
2. Turn on budgets and anomaly detection. AWS Budgets and Cost Anomaly Detection alert you to unexpected spend within days, not at month end.
3. Review monthly. Run a short cost review every month, and a Well-Architected review once or twice a year, so new waste doesn’t pile up.

Where to start

If you only have an afternoon, do checks 1, 2 and 11: delete what is idle, rightsize the biggest instances, and switch on budget alerts. Then plan commitments once usage is stable. If you’d like a second pair of eyes, our free AWS cost review covers all twelve checks and gives you a prioritized plan.