Collecting CPU and memory metrics for AWS Lambda MicroVMs

This article explains how to collect CPU and memory metrics from AWS Lambda MicroVMs using the CloudWatch Agent with telegraf and OpenTelemetry.

• Configure telegraf to emit CPU and memory metrics every 60 seconds from within the MicroVM
• Use OpenTelemetry to process metrics and export them to CloudWatch with custom dimensions from Lambda-injected environment variables
• Install CloudWatch Agent in your Dockerfile and start it as a background process via an entrypoint script
• Ensure the MicroVM execution role has cloudwatch:PutMetricData permissions
• Verify metrics appear in CloudWatch console under the LambdaMicroVms/Application namespace

Monitoring CPU and memory utilization helps right-size MicroVMs, debug service health, and identify scaling signals for production workloads.