Google Associate Cloud Engineer Practice Test 44
Question 430
A production Kubernetes Deployment exists in namespace payments, but kubectl get deployments shows no results because the operator's current namespace is development. How should the operator locate the workload before editing it?
- Delete the development namespace so kubectl automatically switches to whichever cluster is production.
- Change the Cloud Billing account of the cluster; the billing account controls kubectl namespace visibility.
- Verify the kubectl cluster context and specify the payments namespace explicitly when listing or modifying the Deployment.
- Increase node pool size; all workloads become visible in the default namespace once additional nodes join.
Explanation
Correct Answer
C:Verify the kubectl cluster context and specify the payments namespace explicitly when listing or modifying the Deployment. Kubernetes resources can be namespaced. Explicit context and namespace avoid acting on the wrong environment or mistaking an empty list for a missing workload.
Incorrect Answers
A:Delete the development namespace so kubectl automatically switches to whichever cluster is production. Deleting an unrelated namespace is destructive and does not guarantee the intended Kubernetes context or scope is selected.
B:Change the Cloud Billing account of the cluster; the billing account controls kubectl namespace visibility. Billing linkage does not change which namespace a kubectl command queries or which cluster context is active.
D:Increase node pool size; all workloads become visible in the default namespace once additional nodes join. Adding nodes addresses capacity. It does not relocate a Deployment from payments into the default namespace or change the query scope.
Question 431
A GKE HorizontalPodAutoscaler is present but reports unknown CPU utilization because its Deployment does not specify CPU requests. Which workload setting should be fixed?
- Configure a VerticalPodAutoscaler recommendation or supported auto mode for the workload after assessing compatibility with its HPA.
- Move all Pods to the default namespace and assume this repairs scheduling and authorization.
- Grant the GKE node service account project Owner to bypass the workload-specific issue.
- Add realistic per-container CPU resource requests and confirm metrics availability for the autoscaling target.
Explanation
Correct Answer
D:CPU utilization percentage in HPA is based on resource requests; without them the autoscaler cannot calculate the intended percentage correctly.
Incorrect Answers
A:Changing an unrelated Pod setting leaves the stated cloud feature unresolved. The specific requirement concerns horizontal Pod replica scaling.
B:Changing an unrelated Pod setting leaves the stated cloud feature unresolved. The necessary distinction is horizontal Pod replica scaling.
C:Project Owner broadens authorization and is not an appropriate substitute for the named control. The case instead depends on horizontal Pod replica scaling.
Question 432
An Autopilot workload scales vertically but the team keeps manually changing Pod CPU requests without observing recommendations. What Kubernetes control can help?
- Configure a VerticalPodAutoscaler recommendation or supported auto mode for the workload after assessing compatibility with its HPA.
- Add realistic per-container CPU resource requests and confirm metrics availability for the autoscaling target.
- Disable the cluster health checks and treat missing readiness as a successful deployment.
- Recreate the node pool without preserving the necessary workload identity or configuration.
Explanation
Correct Answer
A:Vertical pod autoscaling manages resource requests from usage trends and must account for restart and HPA interaction policies.
Incorrect Answers
B:This changes a neighboring Kubernetes feature, not the stated control. The necessary distinction is the intended autoscaling control.
C:Deploying a separate workload does not automatically configure the existing feature. The case instead depends on the intended autoscaling control.
D:Changing node pools alone cannot substitute for the specified workload-level control. It leaves the intended autoscaling control unaddressed.
Question 433
A GKE application Pod is stuck in CrashLoopBackOff after an image update, although image pull succeeds. Which evidence narrows the problem?
- Inspect previous container logs, termination reason, startup command and readiness/liveness probe configurations.
- Grant read access to the appropriate GKE image-pull identity at the target Artifact Registry repository and validate repository path.
- Increase the Kubernetes Service port without changing the failing Pod or cluster configuration.
- Recreate the node pool without preserving the necessary workload identity or configuration.
Explanation
Correct Answer
A:CrashLoopBackOff indicates recurring container startup failure; successful image pull does not establish application health.
Incorrect Answers
B:This changes a neighboring Kubernetes feature, not the stated control. The case instead depends on the Kubernetes workload or scheduling control.
C:A Kubernetes Service port does not change a Pod’s scheduling, image or identity. It leaves the Kubernetes workload or scheduling control unaddressed.
D:Changing node pools alone cannot substitute for the specified workload-level control. The specific requirement concerns the Kubernetes workload or scheduling control.
Question 434
A production Service shows no ready endpoints even though Pods exist, because readiness probes fail on the wrong path. What should be corrected?
- Align the workload's node affinity or selector with actual supported node labels and eligible capacity.
- Give the attached VM service account project Owner instead of fixing resource configuration.
- Fix the readiness probe path and health behavior before enabling traffic to the Pods.
- Create a snapshot only and assume it immediately changes the live VM capacity.
Explanation
Correct Answer
C:Kubernetes Services route to endpoints considered ready, so a misconfigured readiness probe can withhold healthy backend pods.
Incorrect Answers
A:The proposed setting does not correct the stated machine, disk or scheduling requirement. It leaves the actual compute scheduling or resource condition unaddressed. The appropriate action here is: Fix the readiness probe path and health behavior before enabling traffic to the Pods.
B:Project Owner broadens authorization and is not an appropriate substitute for the named control. The specific requirement concerns the actual compute scheduling or resource condition. The appropriate action here is: Fix the readiness probe path and health behavior before enabling traffic to the Pods.
D:Snapshots are recovery artifacts and do not directly reconfigure a running resource. The necessary distinction is the actual compute scheduling or resource condition. The appropriate action here is: Fix the readiness probe path and health behavior before enabling traffic to the Pods.
Question 435
A cluster has Pods stuck Pending because a node selector references a zone label not present on any available nodes. How should scheduling be repaired?
- Align the workload's node affinity or selector with actual supported node labels and eligible capacity.
- Restore least-privilege registry-reading permission to the new node identity and validate image pulls in a safe node-pool rollout.
- Give the attached VM service account project Owner instead of fixing resource configuration.
- Create a snapshot only and assume it immediately changes the live VM capacity.
Explanation
Correct Answer
A:Kubernetes scheduler constraints are conjunctive; mismatched selectors prevent scheduling even when other nodes have free CPU.
Incorrect Answers
B:The proposed setting does not correct the stated machine, disk or scheduling requirement. The specific requirement concerns the actual compute scheduling or resource condition. The appropriate action here is: Align the workload's node affinity or selector with actual supported node labels and eligible capacity.
C:Project Owner broadens authorization and is not an appropriate substitute for the named control. The necessary distinction is the actual compute scheduling or resource condition. The appropriate action here is: Align the workload's node affinity or selector with actual supported node labels and eligible capacity.
D:Snapshots are recovery artifacts and do not directly reconfigure a running resource. The case instead depends on the actual compute scheduling or resource condition. The appropriate action here is: Align the workload's node affinity or selector with actual supported node labels and eligible capacity.
Question 436
A workload on GKE Standard has Pods tolerating a production-only taint even though they belong to test. What must be tightened?
- Revise hard anti-affinity or provide eligible zonal capacity while retaining genuine availability separation goals.
- Remove overbroad tolerations and apply the intended taint, namespace and workload scheduling policy separation.
- Change only the container image tag and assume a networking or policy failure is repaired.
- Grant the GKE node service account project Owner to bypass the workload-specific issue.
Explanation
Correct Answer
B:Tolerations permit placement on tainted nodes but do not grant true isolation unless eligibility and affinity are also enforced.
Incorrect Answers
A:This changes a neighboring Kubernetes feature, not the stated control. The necessary distinction is the Kubernetes workload or scheduling control.
C:Network settings and service or application controls must be distinguished. The case instead depends on the Kubernetes workload or scheduling control.
D:Project Owner broadens authorization and is not an appropriate substitute for the named control. It leaves the Kubernetes workload or scheduling control unaddressed. The appropriate action here is: Remove overbroad tolerations and apply the intended taint, namespace and workload scheduling policy separation.
Question 437
A team applies a NetworkPolicy to limit Pod-to-Pod egress but traffic remains unrestricted because network policy enforcement is disabled for that cluster mode. What should be verified?
- Set an approved control-plane authorized-network or private endpoint access path under the supported GKE networking mode.
- Grant the GKE node service account project Owner to bypass the workload-specific issue.
- Enable a supported network-policy enforcement path or choose a compatible cluster/network dataplane and test Pod traffic.
- Recreate the node pool without preserving the necessary workload identity or configuration.
Explanation
Correct Answer
C:Declaring Kubernetes NetworkPolicy objects without an enforcing dataplane does not necessarily restrict traffic.
Incorrect Answers
A:Network settings and service or application controls must be distinguished. The case instead depends on the Kubernetes workload or scheduling control. The appropriate action here is: Enable a supported network-policy enforcement path or choose a compatible cluster/network dataplane and test Pod traffic.
B:Project Owner broadens authorization and is not an appropriate substitute for the named control. It leaves the Kubernetes workload or scheduling control unaddressed. The appropriate action here is: Enable a supported network-policy enforcement path or choose a compatible cluster/network dataplane and test Pod traffic.
D:Changing node pools alone cannot substitute for the specified workload-level control. The specific requirement concerns the Kubernetes workload or scheduling control. The appropriate action here is: Enable a supported network-policy enforcement path or choose a compatible cluster/network dataplane and test Pod traffic.
Question 438
A GKE cluster upgrade is scheduled, but a stateful Pod uses an unsupported deprecated Kubernetes API and will fail after upgrade. Which preventive check is needed?
- Review deprecation insights and migrate manifests/controllers to supported APIs before the control-plane upgrade.
- Set an approved control-plane authorized-network or private endpoint access path under the supported GKE networking mode.
- Grant the GKE node service account project Owner to bypass the workload-specific issue.
- Increase the Kubernetes Service port without changing the failing Pod or cluster configuration.
Explanation
Correct Answer
A:Kubernetes API removals can break manifests and controllers independent of whether nodes have enough compute capacity.
Incorrect Answers
B:Network settings and service or application controls must be distinguished. It leaves the Kubernetes workload or scheduling control unaddressed.
C:Project Owner broadens authorization and is not an appropriate substitute for the named control. The specific requirement concerns the Kubernetes workload or scheduling control.
D:A Kubernetes Service port does not change a Pod’s scheduling, image or identity. The necessary distinction is the Kubernetes workload or scheduling control.
Question 439
A PodDisruptionBudget requires all six replicas to remain available during maintenance, blocking an authorized drain. What should be adjusted?
- Review maxSurge, maxUnavailable, quota and template resource availability and choose safe update limits for the serving target.
- Move the VM to a different billing account and assume its scheduling and storage behavior changes.
- Choose a realistic minAvailable or maxUnavailable budget consistent with redundancy and planned maintenance capacity.
- Change the instance display name as a replacement for modifying the actual machine or disk.
Explanation
Correct Answer
C:PodDisruptionBudgets coordinate voluntary disruptions; setting zero permissible disruption can prevent node drains indefinitely.
Incorrect Answers
A:The proposed setting does not correct the stated machine, disk or scheduling requirement. The specific requirement concerns the actual compute scheduling or resource condition. The appropriate action here is: Choose a realistic minAvailable or maxUnavailable budget consistent with redundancy and planned maintenance capacity.
B:Billing changes payment responsibility, not the technical behavior of this workload. The necessary distinction is the actual compute scheduling or resource condition. The appropriate action here is: Choose a realistic minAvailable or maxUnavailable budget consistent with redundancy and planned maintenance capacity.
D:Display names and labels identify resources; they do not enforce their runtime behavior. The case instead depends on the actual compute scheduling or resource condition. The appropriate action here is: Choose a realistic minAvailable or maxUnavailable budget consistent with redundancy and planned maintenance capacity.
Question 440
A team rotates node-pool service-account credentials but new nodes can no longer read authorized Artifact Registry images. What should be verified?
- Grant read access to the appropriate GKE image-pull identity at the target Artifact Registry repository and validate repository path.
- Restore least-privilege registry-reading permission to the new node identity and validate image pulls in a safe node-pool rollout.
- Move the VM to a different billing account and assume its scheduling and storage behavior changes.
- Restart the instance repeatedly while leaving the underlying error or missing dependency in place.
Explanation
Correct Answer
B:GKE node identity and workload identity have different duties; removing node registry access can break container startup.
Incorrect Answers
A:The proposed setting does not correct the stated machine, disk or scheduling requirement. The necessary distinction is the actual compute scheduling or resource condition. The appropriate action here is: Restore least-privilege registry-reading permission to the new node identity and validate image pulls in a safe node-pool rollout.
C:Billing changes payment responsibility, not the technical behavior of this workload. The case instead depends on the actual compute scheduling or resource condition. The appropriate action here is: Restore least-privilege registry-reading permission to the new node identity and validate image pulls in a safe node-pool rollout.
D:The proposed setting does not correct the stated machine, disk or scheduling requirement. It leaves the actual compute scheduling or resource condition unaddressed. The appropriate action here is: Restore least-privilege registry-reading permission to the new node identity and validate image pulls in a safe node-pool rollout.