6.4 Assisting with the Support of Deployed Solutions PROFESSIONAL-CLOUD-ARCHITECT Practice Questions (Page 7)

Questions 31–34

1. 31expert · medium

   A retail company's order-processing service, running on Google Kubernetes Engine (GKE), intermittently fails to process orders. The failures correlate with high CPU usage on the node running the service. The team has already scaled the deployment and added more replicas, but the issue persists. They suspect a specific code path is causing excessive CPU consumption. Which approach would best help identify the root cause?

   • AUse Cloud Profiler to capture CPU profiles from the service and analyze the hot spots.
   • BIncrease the memory limits of the pods to reduce CPU throttling.
   • CRestart the service and monitor the logs for any error messages.
   • DEnable CPU-based horizontal pod autoscaling and set a higher target utilization.
   Reveal Solution Discussion Select an answer first
2. 32application · medium

   A company runs a web application on Compute Engine behind a managed instance group. They want to be alerted if the application becomes unavailable. They have set up a health check. What should they configure to receive notifications?

   • ACreate a Cloud Monitoring alerting policy that uses the health check's uptime metric and sends notifications to a notification channel.
   • BEnable Cloud Logging for the health check and set up a log-based alert.
   • CConfigure the health check to send an email directly when it fails.
   • DSet up a Cloud Scheduler job to run a script that checks the application and sends an email if it is down.
   Reveal Solution Discussion Select an answer first
4. 33expert · medium

   A developer is debugging a serverless application on Cloud Run. The application intermittently fails with a 503 error. The developer has access to Cloud Logging and Cloud Monitoring. The developer suspects the issue is related to cold starts or concurrency limits. Which approach would best confirm the root cause?

   • AIncrease the concurrency limit of the Cloud Run service and see if the errors stop.
   • BExamine the request logs to see the latency and any error messages for the 503 responses.
   • CReview the Cloud Run service's configuration to see if the minimum number of instances is set to zero.
   • DCheck the Cloud Run service's CPU utilization in Cloud Monitoring.
   Reveal Solution Discussion Select an answer first
5. 34expert · medium

   A company uses a blue/green deployment strategy for their application on Google Kubernetes Engine (GKE). After switching traffic to the green version, they notice a critical bug. They need to roll back to the blue version. What is the most efficient way to do this?

   • AScale down the green deployment to zero and scale up the blue deployment.
   • BDelete the green deployment and redeploy the blue version from scratch.
   • CSwitch the traffic back to the blue version by updating the service selector to point to the blue deployment.
   • DUse kubectl rollout undo to revert the green deployment to the previous version.
   Reveal Solution Discussion Select an answer first