3.2 Managing Service Lifecycle PROFESSIONAL-CLOUD-DEVOPS-ENGINEER Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   What is the primary purpose of Dynamic Workload Scheduler?

   • ATo manage quotas and limits
   • BTo schedule flexible workloads during periods of available capacity
   • CTo guarantee immediate execution of all workloads
   • DTo replace autoscaling for all services
   Reveal Solution Discussion Select an answer first
2. 12application · medium

   You deploy a containerized REST API to Cloud Run. The service is CPU-intensive and each request takes about 2 seconds to process. You configure the service with a minimum of 2 instances and a maximum of 20 instances. During a traffic test, you observe that the service scales to 20 instances, but the 95th percentile latency is still high. You check the Cloud Run metrics and see that the average CPU utilization is only 35%. What is the most likely cause of the high latency?

   • AThe container instances are being throttled because the CPU is allocated only during request processing
   • BThe maximum number of instances is too low, preventing the service from scaling enough
   • CThe concurrency setting is too high, causing requests to queue behind each other on each instance
   • DThe minimum number of instances is too low to handle the initial burst of traffic
   Reveal Solution Discussion Select an answer first
4. 13expert · hard

   You run a stateful application on a GKE cluster. The application stores data on regional persistent disks. You configure the Horizontal Pod Autoscaler (HPA) to scale based on CPU utilization at 70%. You also enable the cluster autoscaler. During a load test, you notice that new pods are created, but they remain in 'Pending' state for several minutes. You verify that the cluster has enough total CPU and memory resources. You also verify that the cluster autoscaler is enabled and has a sufficient maximum node count. What is the most likely cause of the pending pods?

   • AThe cluster autoscaler is not able to add nodes in the zones where the persistent disks are located
   • BThe cluster autoscaler has a maximum node count that is too low
   • CThe HPA is scaling based on the wrong metric, so it creates pods that cannot be scheduled
   • DThe persistent volumes are not being created correctly
   Reveal Solution Discussion Select an answer first
5. 14foundation · easy

   Which step is part of the service retirement process?

   • AData migration
   • BInitial capacity planning
   • CCanary analysis
   • DAutoscaling configuration
   Reveal Solution Discussion Select an answer first
6. 15foundation · easy

   What is the primary purpose of Google Cloud quotas?

   • ATo automatically scale resources based on demand
   • BTo reserve capacity for critical workloads
   • CTo limit the amount of a specific resource that a project can use
   • DTo guarantee that resources are always available
   Reveal Solution Discussion Select an answer first