A retail company runs a microservices application on GKE. Users intermittently report that checkout requests are slow, but the frontend team, cart team, and payment team each insist their service is healthy based on their own dashboards. The architect needs to define a troubleshooting process that pinpoints which service or inter-service call is responsible for the latency across a single request as it traverses multiple services. Which approach best establishes this capability?