Question

A company runs a microservices application on ECS Fargate behind an API Gateway. Customers report intermittent high latency on checkout requests, but individual service CloudWatch metrics all appear healthy. The DevOps team has enabled AWS X-Ray on all services and API Gateway. They need to quickly identify which specific downstream dependency is contributing the most latency to the affected checkout requests. Which approach should the team use?

Accepted Answer

Use the X-Ray service map and apply a trace filter expression (e.g., filtering by the checkout URL and high response time) to isolate slow traces, then examine the segment and subsegment timings to pinpoint the bottleneck dependency.

Answer

Create a CloudWatch metric filter on the ECS task logs to extract latency values and build a dashboard that graphs average latency per service.

Answer

Enable CloudWatch Container Insights on the ECS cluster and review the per-container CPU and memory utilization to find the overloaded service.

Answer

Configure CloudWatch Logs Insights to query API Gateway access logs and sort by integration latency to identify the slow endpoint.