A developer needs to diagnose performance issues in an AWS Lambda function interacting with multiple AWS services, without altering the code. What approach should be taken to trace the service interactions and identify the cause?