Question

A company runs a serverless order-processing pipeline consisting of API Gateway, several Lambda functions, an SQS queue, and DynamoDB. During intermittent latency spikes, the operations team cannot determine which component introduces delay, because each service emits logs to separate CloudWatch log groups with no correlated request identifier. They want end-to-end visibility into individual request paths across all components with minimal code changes. Which approach BEST improves observability?

Accepted Answer

Enable AWS X-Ray active tracing on API Gateway and the Lambda functions, and instrument the SQS and DynamoDB calls with the X-Ray SDK to propagate trace context across the pipeline.

Answer

Create a CloudWatch dashboard aggregating the Duration, Errors, and Throttles metrics for each Lambda function and set anomaly-detection alarms.

Answer

Enable CloudWatch Logs Insights and write a query that joins all log groups on the timestamp field to reconstruct request flows.

Answer

Configure CloudTrail data events for Lambda and DynamoDB and forward them to Amazon OpenSearch Service for correlation analysis.