Prepare and test your skills
Prepare and test your skills
Free sample question, answered and explained below. The Professional Cloud DevOps Engineer practice bank has 1,455 more like it. Try 10 free: instant score, no account.
Your organization runs a high-throughput data processing and web application architecture on Google Kubernetes Engine (GKE). During off-peak hours, workload traffic drops substantially; however, FinOps reports indicate that GKE compute costs remain high because underutilized nodes in standard on-demand node pools fail to scale down.
Your investigation reveals the following workload characteristics:
Which strategy should you implement to optimize workload lifecycle management and minimize idle resource costs?
This strategy combines Spot VMs, node pool isolation, and Cluster Autoscaler (CA) annotations to minimize infrastructure expenses while eliminating scale-down blockers across GKE clusters.
"cluster-autoscaler.kubernetes.io/safe-to-evict": "true" annotation explicitly instructs CA that pods utilizing local storage can be safely evicted and rescheduled elsewhere during scale-down operations.This architecture directly addresses the root causes of scale-down failures without compromising service reliability. It matches workload fault tolerance with cost-efficient compute (Spot VMs) while giving the Cluster Autoscaler the permissions and workload separation it requires to aggressively eliminate idle infrastructure waste.
This option configures Horizontal Pod Autoscaler (HPA) and Vertical Pod Autoscaler (VPA) concurrently on the exact same resource metric (CPU utilization) while deleting all Pod Disruption Budgets (PDBs) to force node scale-downs.
VPA was configured strictly in recommendation mode (off) to right-size pod requests while HPA managed dynamic traffic scaling, accompanied by properly defined PDBs.
This option disables automated infrastructure elasticity by switching from Cluster Autoscaler to manual node allocation, provisioning oversized high-memory VMs, and removing pod-level CPU and memory requests.
The workload required a strictly static, non-bursting, single-tenant bare-metal topology where autoscaling latency could not be tolerated under any circumstance.
This approach sets the "cluster-autoscaler.kubernetes.io/safe-to-evict": "false" annotation across all Kubernetes pods and attempts to shift batch background tasks to Cloud Run with request-based billing.
"safe-to-evict": "false" explicitly forbids the Cluster Autoscaler from evicting any pod during node drains. This worsens idle node utilization and increases GKE compute spend.safe-to-evict: false turns every node hosting an annotated pod into an unremovable instance.You needed to guarantee that mission-critical, un-checkpointed stateful pods on GKE were never evicted or terminated by automated maintenance and autoscaling events.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated