Google Professional Cloud DevOps Engineer Question 154
Select 3Google Cloud PlatformYour organization runs a web application on Google Kubernetes Engine (GKE). Recently, users have reported intermittent high latency and timeouts while accessing the application. You suspect an issue with the backend pods. Which actions should you take to diagnose and resolve the issue?
- A
Use Stackdriver Profiler to analyze code-level performance issues in the backend application.
- B
Inspect the pod logs using
kubectl logsto check for errors or warnings in the backend containers. - C
Use Cloud Monitoring to evaluate the CPU and memory utilization of backend pods.
- D
Verify the pod's readiness and liveness probes configuration to ensure proper health checks are in place.
- E
Scale up the number of backend pods by increasing the replica count in the deployment configuration.
Show answer and explanation
Correct answers: B, C, D
Explanation
When users report high latency or timeouts in a GKE-based application, diagnosing backend pod issues requires a systematic approach. Logs, resource utilization metrics, and health check configurations are key areas to investigate. While scaling might mitigate symptoms, resolving the underlying issue is critical to ensure long-term stability and performance.
- A. Incorrect.
Stackdriver Profiler (now part of Cloud Profiler) is useful for analyzing code-level performance issues, but it does not directly help in diagnosing runtime issues like pod crashes, excessive resource usage, or misconfigured health checks.
- B. Correct.
kubectl logsallows you to inspect logs from the backend pods, which can provide insight into runtime errors or warnings causing the high latency or timeouts. - C. Correct.
Cloud Monitoring can help identify resource bottlenecks such as high CPU or memory utilization, which might lead to degraded performance of the backend pods.
- D. Correct.
Misconfigured readiness or liveness probes could lead to unhealthy pods serving traffic or being terminated prematurely, contributing to intermittent latency or timeouts.
- E. Incorrect.
Scaling up the number of backend pods might temporarily alleviate latency issues, but it does not address the root cause of the problem. Diagnosing the issue should take precedence before scaling.