Question

A CloudOps engineer manages a fleet of EC2-based web servers behind an Application Load Balancer. Business stakeholders want a single CloudWatch metric that reflects overall application availability, calculated from the ratio of successful (2xx/3xx) responses to total requests over each 5-minute window. The ALB access logs are delivered to S3, and a Lambda function already parses them. What is the MOST appropriate way to expose this availability figure so it can be alarmed on and shown on a dashboard?

Accepted Answer

Have the Lambda function calculate the ratio and publish it to CloudWatch as a custom metric using PutMetricData with a defined namespace and dimensions

Answer

Create a CloudWatch metric math expression on the built-in ALB HTTPCode_Target_2XX_Count metric and store the result as a new AWS/ApplicationELB metric

Answer

Configure a CloudWatch Logs metric filter on the S3 access logs to count successful responses and emit it as a metric

Answer

Enable CloudWatch anomaly detection on the RequestCount metric and use the anomaly band as the availability indicator