What to measure before resizing an EC2 instance

By Yunzhihui Cloud Field Notes · Wed Nov 18 2026

A slow API and an oversized batch worker can both make an EC2 instance look like a candidate for resizing. Start by identifying what the workload needs to improve, then compare the same workload before and after the change.

Record one representative window

Capture a busy period and a quiet period. Record request rate, p95 response latency and error rate alongside CPU utilization. Check memory and swap with an OS-level collector such as the CloudWatch agent, since standard EC2 metrics do not include guest memory usage. If you run a burstable instance, also check CPU credit balance before interpreting a CPU average.

Separate compute from storage and network

A queue that grows during backups can point to storage contention. Compare volume latency and throughput with application latency, then check both the EBS volume and instance limits. For a network-heavy service, compare concurrent request rate and network throughput. Change one variable at a time so the result is interpretable.

Define the rollback before changing size

Check compatibility for the target instance type, including architecture and storage. Schedule downtime if the resize requires stopping the instance, identify any instance-store data that must be copied elsewhere first, and document how to return to the previous configuration. After the change, compare the same traffic window and review total cost, including storage and data transfer.

For a broader change plan, see our cloud migration checklist. This article is from Yunzhihui Cloud and was drafted with AI assistance.

Read on NewsPlatform · More stories