An AWS cost optimisation checklist for growing products

Cloud bills tend to grow quietly: a test environment left running, an oversized database, logs kept forever. The good news is that most products can cut waste significantly without touching reliability. Use this checklist as a quarterly routine.

1. Get visibility first

• Turn on Cost Explorer and set budgets with alerts.
• Tag every resource by environment, team and product, and enforce tags in your infrastructure code.
• Review the top ten cost lines every month — that is usually where most of the spend sits.

2. Remove what you don’t use

• Delete unattached volumes, old snapshots and unused load balancers and IP addresses.
• Schedule non-production environments to shut down outside working hours.
• Clean up abandoned experiments and forgotten regions.

3. Right-size compute and databases

• Check CPU and memory utilisation; many instances run far below capacity.
• Move to newer instance generations, including Graviton (ARM) where your stack supports it.
• Use auto scaling so capacity follows demand.

4. Use the right pricing model

• Cover steady baseline usage with Savings Plans or reserved capacity.
• Use Spot capacity for fault-tolerant workloads such as batch jobs and CI runners.
• Consider serverless options for spiky, low-average workloads.

5. Optimise storage

• Apply S3 lifecycle policies to move older data to cheaper tiers or expire it.
• Set log retention periods instead of keeping logs forever.
• Compress and partition analytical data.

6. Watch data transfer

• Serve static assets through a CDN.
• Keep chatty services in the same region and availability zone where appropriate.
• Use VPC endpoints to avoid unnecessary NAT gateway charges.

7. Build cost into engineering

Make cost visible in pull requests and dashboards, and include it in architecture reviews. When engineers can see the price of their choices, costs stay under control.

Need a second pair of eyes?

AkMo designs and runs the AWS infrastructure behind our own products. Our cloud and DevOps team can review your architecture and identify savings — get in touch.