During an incident, calls from a checkout service to a fraud-scoring service in GKE begin timing out, and failing calls keep piling up. The team needs the failures to be cut off quickly without changing application code. Which control should they add?
Show answer and explanation
Answer: D
Circuit breaking is configured in the mesh. Connection pool limits cap concurrent requests and pending work to a service, and outlier detection temporarily ejects endpoints that keep failing. When the fraud-scoring service degrades, the mesh fails calls fast instead of queuing them, which stops the cascade. Cloud Service Mesh, based on Istio, applies these traffic policies without changes to the checkout code.
Why the other options are wrong
- A. Unlimited retries with no timeout add load to a service that is already failing and tie up checkout threads, which tends to amplify the outage.
- B. More checkout replicas add callers and send more requests to the struggling fraud service, so the failing calls keep piling up.
- C. A public load balancer without health checks keeps sending traffic to unhealthy backends, and nothing limits outstanding connections.