Google Associate Cloud Engineer Practice Test 9
Question 80
A user successfully signs in through the enterprise IdP but a Google Cloud session is denied due to a missing workforce federation mapping. What component should be inspected?
- Grant all federated users organization Owner so the provider mapping is optional.
- Update a project IAM binding to refer to an unrecognized workforce subject string without fixing mapping.
- Inspect the workforce provider’s claim mappings and the principal identifiers used by IAM bindings.
- Create a user-managed service account key and have employees sign in with it.
Explanation
Correct Answer
C:Successful IdP authentication is only part of federation. Google Cloud must map the IdP’s assertions into the expected workforce subject or attribute identifiers and match those principals to resource IAM bindings.
Incorrect Answers
A:Broad resource permissions cannot replace a valid mapping from external assertions to workforce principals.
B:A principal that does not match the provider's actual mapped identity will not fix the authorization mismatch.
D:Service-account keys are workload credentials and do not repair a missing interactive-user mapping.
Question 81
A security review finds that a Viewer role on project payroll is bound to every identity in workforce pool external-contractors. Only contract team alpha needs access. Which least-privilege change is appropriate?
- Convert the Viewer grant to Owner at the same pool-wide principalSet; the larger role makes team filtering automatic.
- Disable Cloud Billing on payroll; suspended billing removes unrelated principals from IAM allow policies.
- Add another Viewer binding for team alpha while retaining the pool-wide binding unchanged.
- Replace the pool-wide grant with a principalSet matching a trusted group or mapped team attribute for alpha.
Explanation
Correct Answer
D:Replace the pool-wide grant with a principalSet matching a trusted group or mapped team attribute for alpha. Google supports group- and attribute-scoped workforce principalSets. Replacing the broad principalSet restricts this project role to the intended team.
Incorrect Answers
A:Convert the Viewer grant to Owner at the same pool-wide principalSet; the larger role makes team filtering automatic. Owner increases privilege and still includes the whole pool, so it compounds both the scope and permission problems.
B:Disable Cloud Billing on payroll; suspended billing removes unrelated principals from IAM allow policies. Billing state does not change membership of an IAM principalSet. Disabling billing also disrupts workloads unnecessarily.
C:Add another Viewer binding for team alpha while retaining the pool-wide binding unchanged. IAM allow bindings are additive. An additional narrow grant cannot remove the overly broad existing workforce-pool grant.
Question 82
A company federates employees through an external identity provider, but successive logins create duplicate workforce principals because the subject mapping uses changing display names. What must be stabilized?
- Map a stable provider-issued user identifier into the workforce subject attribute.
- Use the employee’s current display name as the only workforce subject on every sign-in.
- Assign every session a different random workforce subject and grant each one Owner.
- Disable assertion validation and identify users solely by the source IP address.
Explanation
Correct Answer
A:A subject derived from a durable identifier remains stable when a user changes their visible name. Verify that the issuer’s chosen claim stays constant across sessions, then map it consistently so IAM sees the same workforce principal.
Incorrect Answers
B:Display names can change and may not be unique, creating identity inconsistency.
C:Randomizing the subject breaks persistent IAM bindings and makes access review impractical.
D:Source IPs do not provide reliable principal identity and do not substitute for trusted assertions.
Question 83
A new project lacks Cloud Run API activation in its effective Service Usage policy, but the release pipeline already has the correct Cloud Run deployment role. What prerequisite must be completed?
- Enable Cloud Run in the project using approved Service Usage authorization or an eligible inherited enablement policy before deploying.
- Grant an organization-wide Owner role to the deployment account because deployment permissions include all service activation by default.
- Resize the project's default VPC subnet because service activation depends on the subnet address range.
- Add a Cloud Billing budget alert because any project with cost alerts automatically enables Cloud Run.
Explanation
Correct Answer
A:Enable Cloud Run in the project using approved Service Usage authorization or an eligible inherited enablement policy before deploying. Cloud Run deployment authorization and service activation are independent. The effective service state must include Cloud Run; project-level enablement or the supported hierarchical activation Preview can provide that state.
Incorrect Answers
B:Grant an organization-wide Owner role to the deployment account because deployment permissions include all service activation by default. A broad Owner role is an excessive response to a separate Service Usage prerequisite, and a role grant does not itself enable the Cloud Run API.
C:Resize the project's default VPC subnet because service activation depends on the subnet address range. VPC address space does not determine whether a product API is enabled for the project. Subnet resizing cannot correct a Service Usage disabled state.
D:Add a Cloud Billing budget alert because any project with cost alerts automatically enables Cloud Run. Billing alerts track spending, not API activation; their existence cannot make a disabled Cloud Run service available.
Question 84
Container API usage works in a development project, but production is in a different folder and its effective Service Usage policy does not enable the Container API. What should the platform team do before creating GKE clusters in production?
- Assume that enabling an API anywhere in the company automatically enables it in every other project.
- Move the GKE cluster into the development project, bypassing production governance for the release.
- Enable the Container API through an approved production project or eligible ancestor policy and verify the effective state.
- Grant network firewall administrator in the production project and expect that role to activate every Google Cloud API.
Explanation
Correct Answer
C:Enable the Container API through an approved production project or eligible ancestor policy and verify the effective state. A production project must have an effective enabled Container API before creating GKE clusters. Service Usage supports local enablement and, in Preview, inherited enablement policies.
Incorrect Answers
A:Assume that enabling an API anywhere in the company automatically enables it in every other project. A project's service state is not inferred simply from another unrelated project. Hierarchical enablement applies only when an appropriate ancestor policy actually enables that service.
B:Move the GKE cluster into the development project, bypassing production governance for the release. Moving the workload to development avoids the stated production requirement and violates the intended governance boundary rather than preparing production correctly.
D:Grant network firewall administrator in the production project and expect that role to activate every Google Cloud API. Firewall administration deals with network security, not the Service Usage enablement state for the GKE Container API.
Question 85
A release automation has permission to deploy Cloud Run services, but enabling the currently disabled Cloud Run API returns PERMISSION_DENIED. Which distinct project permission is needed?
- Compute Engine instanceAdmin permission, because all Google APIs are enabled as part of VM administration.
- The Service Usage enable permission (serviceusage.services.enable), provided through an approved role to the authorized operator.
- Billing Account Viewer, because seeing charges is sufficient to enable any disabled project service.
- roles/logging.viewer, because API activation is a Cloud Logging data-read operation.
Explanation
Correct Answer
B:The Service Usage enable permission (serviceusage.services.enable), provided through an approved role to the authorized operator. The documented Service Usage enable permission gates service enablement. The CI identity or approved administrator needs that specific authorization.
Incorrect Answers
A:Compute Engine instanceAdmin permission, because all Google APIs are enabled as part of VM administration. Compute Engine instance administration does not authorize service activation. The deployment role and Service Usage privileges are separate.
C:Billing Account Viewer, because seeing charges is sufficient to enable any disabled project service. Billing visibility does not change whether a project may enable an API; an active billing account and Service Usage permission are distinct.
D:Roles/logging.viewer, because API activation is a Cloud Logging data-read operation. Viewing log entries cannot modify the enabled-service set, so this role cannot resolve a service activation permission denial.
Question 86
A project team wants to minimize unintended API exposure and bills by identifying services currently enabled. Which inventory should they examine?
- List the project’s enabled APIs using Service Usage and review dependencies before disabling any.
- Search only the Cloud Billing invoice for API names and treat every zero-cost service as disabled.
- Read the VM instance list because it contains all enabled project APIs.
- Disable every service not mentioned in the project’s display-name description.
Explanation
Correct Answer
A:Service Usage reports which services are enabled for the selected project. Compare that inventory with actual workload dependencies before removing unused APIs; the enablement list itself does not prove that a service costs money or is safe to disable.
Incorrect Answers
B:Invoices describe charges rather than serving as a complete service enablement inventory.
C:An instance list covers only VM resources, not all services that may be enabled.
D:A display name is not a dependency inventory and may omit critical background services.
Question 87
An engineer disables one API during cleanup and accidentally breaks a scheduled data processing job. What process would have prevented it?
- Verify that the API is disabled in the console, but skip checking the jobs that use its endpoint.
- Increase the project’s billing budget before disabling the API.
- Delete the scheduled job first and assume it is no longer required.
- Inventory workload dependencies, test the change and prepare a rollback before disabling the service.
Explanation
Correct Answer
D:A scheduled pipeline can depend on a service that looks unused during a narrow observation window. Identify actual consumers, test the disablement impact, coordinate the change and retain a reversible recovery path.
Incorrect Answers
A:Console state alone does not show whether production pipelines still depend on a service.
B:Higher spending limits do not prevent a service-disabled error.
C:Removing the job destroys the dependent workload rather than making the proposed cleanup safe.
Question 88
A project reaches its Compute Engine regional CPU limit while another region has spare quota. What should the operator check before requesting more capacity?
- Assume the spare CPU quota in another region can be borrowed without changing the deployment region.
- Request more quota in the region that already has spare capacity instead of the region named in the error.
- Check the exact regional quota metric, requested machine type and target region capacity.
- Increase the Cloud Billing alert threshold because alerts are the same as quota limits.
Explanation
Correct Answer
C:The failed request names a regional CPU limit. Confirm the relevant quota metric and target region, then compare the desired resource type and available capacity before deciding whether to request an increase or deploy elsewhere.
Incorrect Answers
A:Quota is often scoped to a particular project and region; unused allowance elsewhere does not automatically transfer.
B:Quota is scoped to the actual failing region; increasing unrelated headroom would not resolve the requested deployment.
D:A budget alert does not grant additional CPUs or change provisioning limits.
Question 89
A batch controller retries Compute Engine VM creation after repeated RESOURCE_EXHAUSTED messages naming the regional CPU quota. No CPUs have been freed. Which next step actually addresses this failure?
- Retry every second with the same instance size until a quota increase occurs automatically.
- Increase the VM boot-disk size because CPU quotas are calculated from persistent-disk capacity.
- Inspect the exact regional CPU quota and usage, then reduce planned concurrency or request an authorized limit increase.
- Change the instance network tag; tag updates automatically recalculate available CPU quota.
Explanation
Correct Answer
C:Inspect the exact regional CPU quota and usage, then reduce planned concurrency or request an authorized limit increase. Quota is evaluated by metric and scope. Adjusting the batch to fit the regional allowance or obtaining more approved quota directly addresses this exhausted metric.
Incorrect Answers
A:Retry every second with the same instance size until a quota increase occurs automatically. Identical repeated requests do not raise the granted regional CPU limit. Without reducing consumption or obtaining a quota increase, the documented rejection persists.
B:Increase the VM boot-disk size because CPU quotas are calculated from persistent-disk capacity. Boot disk capacity is a different resource dimension. Increasing it does not change the CPU quota metric identified in the error.
D:Change the instance network tag; tag updates automatically recalculate available CPU quota. Network tags influence firewall targeting, not quota enforcement. Renaming the VM networking metadata does not release regional CPU allocation.
Question 90
A team needs additional GPU allocation for a short project, but the resource count would exceed its granted quota. What must precede the launch?
- Request additional CPUs only and ignore GPU family quotas.
- Check the GPU-family quota and zonal availability, then request appropriate quota or choose supported capacity.
- Choose a different GPU model without checking that application's accelerator and driver requirements.
- Move the project to a different folder so GPUs no longer count against quotas.
Explanation
Correct Answer
B:GPU launches require both appropriate quota and actual availability in the intended location. Verify the GPU model and quota scope, request an approved adjustment when needed, and select an available supported zone or accelerator if capacity is constrained.
Incorrect Answers
A:GPU allocations can have resource-specific limits separate from CPU allowances.
C:Changing hardware blindly may make the workload incompatible even if quota becomes available.
D:Folder placement cannot itself create physical GPU capacity or remove the project’s enforced quotas.