Question 390
An application needs more Pod replicas when CPU utilization rises. Which configuration best meets the requirement?
- Enable or adjust cluster autoscaler for the eligible node pool
- Enable or adjust GKE cluster autoscaling on a suitable node pool
- Configure a Horizontal Pod Autoscaler for the workload
- Adjust Pod resource requests to meet Autopilot constraints
- Adjust resource requests and limits or use supported Vertical Pod Autoscaler recommendations
Explanation
Correct Answer
C:HPA adjusts Pod replica counts using selected metrics. It does not increase the CPU request of an individual existing Pod.
Incorrect Answers
A:The result of this choice would be to scale Kubernetes worker node count. The requested outcome, to respond to service load by adding Pods, is handled at a different layer or through another operation.
B:The choice is aimed at a different purpose: to provide capacity after workload scale-out; Even if it can scale GKE nodes for pending replicas, that would not accomplish the stated requirement to respond to service load by adding Pods.
D:Administrators use it to fit Pod requests to Autopilot rules, rather than to respond to service load by adding Pods. Applying it would change the wrong component of this deployment.
E:While this can right-size container memory allocations, the present task requires an action to respond to service load by adding Pods. This operation does not supply the behavior described in the requirements.
Question 391
A deployment's containers consistently request too little memory and receive OOM kills. What should be configured?
- Configure a Horizontal Pod Autoscaler for the workload
- Adjust Pod resource requests to meet Autopilot constraints
- Enable or adjust cluster autoscaler for the eligible node pool
- Adjust resource requests and limits or use supported Vertical Pod Autoscaler recommendations
- Use the supported GKE node pool upgrade process
Explanation
Correct Answer
D:Pod resource sizing determines scheduling and enforcement at container level. Increasing node count alone does not fix an undersized memory limit.
Incorrect Answers
A:The tool can scale Kubernetes workload replicas, but it cannot be substituted for the capability needed to prevent OOM failures caused by Pod sizing. Its operational effect is not equivalent.
B:This choice is useful to fit Pod requests to Autopilot rules. The stated task is to prevent OOM failures caused by Pod sizing, which requires a different mechanism or permission scope.
C:Select this when the goal is to scale Kubernetes worker node count. Here the decisive requirement is to prevent OOM failures caused by Pod sizing, and that outcome is not achieved by this action.
E:This alternative addresses gke node pool management by allowing a team to upgrade an existing node pool; The incident, however, requires the different action to prevent OOM failures caused by Pod sizing, so its central configuration remains unchanged.
Question 392
An Autopilot deployment requests CPU and memory combinations outside supported bounds. Which action is most appropriate?
- Enable or adjust cluster autoscaler for the eligible node pool
- Adjust Pod resource requests to meet Autopilot constraints
- Configure a Horizontal Pod Autoscaler for the workload
- Add a node pool with the required supported machine configuration
- Adjust resource requests and limits or use supported Vertical Pod Autoscaler recommendations
Explanation
Correct Answer
B:GKE Autopilot applies compute class and resource request rules. Unsupported request shapes can be modified or rejected according to platform behavior.
Incorrect Answers
A:Its purpose is to scale Kubernetes worker node count. That distinction matters: the scenario calls for an action to resolve incompatible Autopilot resource settings, not this separate configuration task.
C:This setting would scale Kubernetes workload replicas. That behavior does not establish the requested ability to resolve incompatible Autopilot resource settings; it addresses an adjacent concern instead.
D:The technical effect is to add specialized Kubernetes worker capacity, not to resolve incompatible Autopilot resource settings; That distinction matters here because the first operation serves separate memory-heavy nodes, rather than the scenario's required result.
E:This approach can right-size container memory allocations. The requirement is to resolve incompatible Autopilot resource settings, so it controls a different part of the workflow and leaves that result unaddressed.
Question 393
A GKE Deployment with several replicas must roll out a new container image while keeping serving capacity whenever sufficient node resources are available. Which TWO settings help that rollout? (Choose TWO.)
- Use a RollingUpdate strategy with a maxUnavailable setting of zero for the Deployment
- Configure a meaningful readiness probe so new Pods receive traffic only when ready
- Use the Recreate rollout strategy, which terminates all old Pods before new ones are ready
- Configure liveness to kill every new Pod immediately at container startup
Explanation
Correct Answers
A:Setting maxUnavailable to zero aims to retain the configured number of available replicas during updates. Readiness and actual available cluster capacity still affect achievable availability.
B:Readiness determines whether a Pod is eligible for Service traffic. A correct probe prevents the rollout from treating containers as available before they can handle requests.
Incorrect Answers
C:Recreate intentionally stops the prior replicas before replacing them. It conflicts with maintaining serving capacity during a normal rolling release.
D:An aggressive liveness probe causes unnecessary restarts and can worsen availability. It is not a strategy for gradually admitting healthy new replicas.
Question 394
A Deployment should restart its Pods to pick up configuration changes without editing the image tag. What is the best next step?
- Use kubectl rollout restart for the Deployment
- Use kubectl rollout status for the Deployment
- Run kubectl scale on the desired controller
- Run kubectl rollout undo on the Deployment
Explanation
Correct Answer
A:Rollout restart updates the Pod template annotation to trigger replacement. It does not change external Secret values by itself.
Incorrect Answers
B:This setting would check a Kubernetes rollout's completion. That behavior does not establish the requested ability to refresh running Kubernetes workloads; it addresses an adjacent concern instead.
C:This choice is useful to manually adjust Kubernetes replicas. The stated task is to refresh running Kubernetes workloads, which requires a different mechanism or permission scope.
D:Using this action would roll back a Kubernetes Deployment, not refresh running Kubernetes workloads. Its function is different from the specific change described in the scenario.
Question 395
An operator wants to change the replica count to a fixed number for a short-lived demand spike. Which configuration best meets the requirement?
- Run kubectl scale on the desired controller
- Use kubectl rollout status for the Deployment
- Run kubectl rollout undo on the Deployment
- Use kubectl rollout restart for the Deployment
- Shift Cloud Run traffic allocation back to the healthy prior revision
Explanation
Correct Answer
A:Scaling changes the requested number of Pods for a supported workload controller. An active HPA may subsequently override manual replica changes.
Incorrect Answers
B:This mechanism is intended to check a Kubernetes rollout's completion. It does not implement the required action to set a temporary Deployment size, even though both are relevant to the same general service.
C:The result of this choice would be to roll back a Kubernetes Deployment. The requested outcome, to set a temporary Deployment size, is handled at a different layer or through another operation.
D:Administrators use it to restart Deployment Pods in a controlled rollout, rather than to set a temporary Deployment size. Applying it would change the wrong component of this deployment.
E:The technical effect is to roll back active Cloud Run request routing, not to set a temporary Deployment size; That distinction matters here because the first operation serves recover from a bad serverless release, rather than the scenario's required result.
Question 396
A team wants to verify that a Deployment rollout has finished successfully. What should be configured?
- Use kubectl rollout status for the Deployment
- Configure a Horizontal Pod Autoscaler for the workload
- Use kubectl rollout restart for the Deployment
- Run kubectl rollout undo on the Deployment
- Run kubectl scale on the desired controller
Explanation
Correct Answer
A:Rollout status observes progress toward the updated desired state. It can reveal when a deployment is stuck rather than assuming apply completed the workload change.
Incorrect Answers
B:Choosing this would scale Kubernetes workload replicas, an appropriate action for cluster autoscaling and provisioning operations; It cannot take the place of the separate step needed to verify a release has converged in this specific situation.
C:Using this action would restart Deployment Pods in a controlled rollout, not verify a release has converged. Its function is different from the specific change described in the scenario.
D:This choice is useful to roll back a Kubernetes Deployment. The stated task is to verify a release has converged, which requires a different mechanism or permission scope.
E:The tool can manually adjust Kubernetes replicas, but it cannot be substituted for the capability needed to verify a release has converged. Its operational effect is not equivalent.
Question 397
A stateful Pod needs block storage that persists beyond Pod restarts. Which approach should be selected?
- Select an appropriate Kubernetes StorageClass in its claim
- Use a Kubernetes Secret or approved external secret integration
- Create a PersistentVolumeClaim using a supported StorageClass
- Mount the ConfigMap or expose its values through the Pod spec
Explanation
Correct Answer
C:PVCs request persistent storage through Kubernetes storage integration. Pod local ephemeral storage is not equivalent to a persistent volume.
Incorrect Answers
A:This approach can choose a Kubernetes persistent storage class. The requirement is to preserve data across Pod replacement, so it controls a different part of the workflow and leaves that result unaddressed.
B:Administrators use it to supply sensitive values to Kubernetes Pods, rather than to preserve data across Pod replacement. Applying it would change the wrong component of this deployment.
D:Its purpose is to inject non-secret Kubernetes configuration. That distinction matters: the scenario calls for an action to preserve data across Pod replacement, not this separate configuration task.
Question 398
A workload needs a particular persistent storage implementation provisioned automatically. Which method addresses the task?
- Use kubectl rollout restart for the Deployment
- Select an appropriate Kubernetes StorageClass in its claim
- Use a Kubernetes Secret or approved external secret integration
- Mount the ConfigMap or expose its values through the Pod spec
- Create a PersistentVolumeClaim using a supported StorageClass
Explanation
Correct Answer
B:StorageClasses define provisioner and class parameters for dynamic volumes. The claim's storage class influences resulting disk behavior.
Incorrect Answers
A:The technical effect is to restart Deployment Pods in a controlled rollout, not to provision a matching storage backend; That distinction matters here because the first operation serves refresh running Kubernetes workloads, rather than the scenario's required result.
C:It can supply sensitive values to Kubernetes Pods; however, it does not provide the specific operation to provision a matching storage backend. The two controls serve different purposes.
D:This mechanism is intended to inject non-secret Kubernetes configuration. It does not implement the required action to provision a matching storage backend, even though both are relevant to the same general service.
E:The tool can request persistent Kubernetes storage, but it cannot be substituted for the capability needed to provision a matching storage backend. Its operational effect is not equivalent.
Question 399
A container needs a ConfigMap's non-secret values at runtime. What is the best next step?
- Use a Kubernetes Secret or approved external secret integration
- Create a PersistentVolumeClaim using a supported StorageClass
- Select an appropriate Kubernetes StorageClass in its claim
- Mount the ConfigMap or expose its values through the Pod spec
- Run kubectl logs for the selected Pod or container
Explanation
Correct Answer
D:ConfigMaps distribute ordinary application configuration to Pods. Sensitive credentials should be handled with appropriate secret-management controls.
Incorrect Answers
A:This setting would supply sensitive values to Kubernetes Pods. That behavior does not establish the requested ability to provide runtime feature settings to Pods; it addresses an adjacent concern instead.
B:Using this action would request persistent Kubernetes storage, not provide runtime feature settings to Pods. Its function is different from the specific change described in the scenario.
C:This approach can choose a Kubernetes persistent storage class. The requirement is to provide runtime feature settings to Pods, so it controls a different part of the workflow and leaves that result unaddressed.
E:Choosing this would inspect a Pod container's emitted logs, an appropriate action for gke service observability and operations operations; It cannot take the place of the separate step needed to provide runtime feature settings to Pods in this specific situation.
Question 400
An application must consume a password without placing it directly in a Deployment manifest. Which configuration best meets the requirement?
- Select an appropriate Kubernetes StorageClass in its claim
- Mount the ConfigMap or expose its values through the Pod spec
- Use a Kubernetes Secret or approved external secret integration
- Create a PersistentVolumeClaim using a supported StorageClass
- Run kubectl rollout undo on the Deployment
Explanation
Correct Answer
C:Secrets provide a Kubernetes interface for sensitive values with access-control requirements. They still need encryption and RBAC practices appropriate to risk.
Incorrect Answers
A:Select this when the goal is to choose a Kubernetes persistent storage class. Here the decisive requirement is to avoid hardcoded manifest passwords, and that outcome is not achieved by this action.
B:This mechanism is intended to inject non-secret Kubernetes configuration. It does not implement the required action to avoid hardcoded manifest passwords, even though both are relevant to the same general service.
D:Administrators use it to request persistent Kubernetes storage, rather than to avoid hardcoded manifest passwords. Applying it would change the wrong component of this deployment.
E:This proposal can roll back a Kubernetes Deployment, but the failure or goal described here calls for a way to avoid hardcoded manifest passwords; It changes the wrong control or scope for the requested outcome.