Question

A company runs a serverless data-processing pipeline built on hundreds of AWS Lambda functions. The finance team reports that Lambda charges have grown steadily, and engineers suspect many functions are configured with more memory than they actually need. The team wants a low-effort, data-driven way to identify functions whose memory allocation could be reduced to lower cost without impacting performance. Which approach best meets this requirement?

Accepted Answer

Enable AWS Compute Optimizer and review its Lambda function memory-sizing recommendations.

Answer

Purchase a Compute Savings Plan to reduce the per-invocation cost of all Lambda functions.

Answer

Manually lower the memory setting on every function to the minimum of 128 MB and monitor for errors.

Answer

Migrate the most expensive functions to EC2 Spot Instances managed by an Auto Scaling group.