You have an application deployed on Google Kubernetes Engine that makes multiple service calls per request and is experiencing slow response times. How can you determine which specific downstream service(s) are causing the performance bottleneck?