Google Associate Cloud Engineer Practice Test 73
Question 720
A service account that was deleted months ago appears in an old IAM policy as a deleted principal. What should be cleaned up?
- Rebind resource IAM to the current authorized workload identity pool and Kubernetes principal identifiers.
- Grant a new billing role instead of adjusting the actual token or impersonation permission.
- Allow tokens from any issuer without verifying audience, signature or mapped principal.
- Remove obsolete deleted-principal bindings and replace them only with intentionally approved active identities.
Explanation
Correct Answer
D:IAM policies can retain historical references; leaving stale bindings is misleading and doesn't restore authorization to a new identity.
Incorrect Answers
A:The suggestion does not establish the required issuer, principal or permission boundary. The necessary distinction is the correct runtime principal and its bindings. The appropriate action here is: Remove obsolete deleted-principal bindings and replace them only with intentionally approved active identities.
B:The suggestion does not establish the required issuer, principal or permission boundary. The case instead depends on the correct runtime principal and its bindings. The appropriate action here is: Remove obsolete deleted-principal bindings and replace them only with intentionally approved active identities.
C:The suggestion does not establish the required issuer, principal or permission boundary. It leaves the correct runtime principal and its bindings unaddressed. The appropriate action here is: Remove obsolete deleted-principal bindings and replace them only with intentionally approved active identities.
Question 721
An administrator creates a service account directly in a production project without linking it to an owning application or rotation policy. What should the creation workflow require?
- Record workload owner, intended permissions, identity lifetime and credential strategy in the approved service-account inventory.
- Use approved service-account impersonation for the deployment commands and record the originating caller in audit evidence.
- Grant Organization Owner to the caller rather than validating the intended identity binding.
- Issue a long-lived service-account JSON key instead of repairing short-lived workload credentials.
Explanation
Correct Answer
A:Account metadata and ownership are needed to retire orphaned service identities safely and to perform later access reviews.
Incorrect Answers
B:Deploying a separate workload does not automatically configure the existing feature. The case instead depends on the correct runtime principal and its bindings.
C:Organization Owner is excessive privilege and cannot repair an unrelated configuration fault. It leaves the correct runtime principal and its bindings unaddressed.
D:The suggestion does not establish the required issuer, principal or permission boundary. The specific requirement concerns the correct runtime principal and its bindings. The appropriate action here is: Record workload owner, intended permissions, identity lifetime and credential strategy in the approved service-account inventory.
Question 722
An internal batch job uses an attached service account whose roles are sufficient, but an API call fails because a VM's legacy OAuth access scopes restrict the token. What second permission layer must be checked?
- Use the appropriate OAuth access token for Google Cloud API scopes rather than a service-targeted OIDC ID token.
- Rename the service account but leave the old principal and role bindings unchanged.
- Verify the instance's OAuth access scopes alongside IAM permissions and use compatible approved scopes for the workload.
- Disable service-account audit logging and assume the credential issue is resolved.
Explanation
Correct Answer
C:On some Compute Engine authentication paths, access scopes can constrain token operations independently of IAM roles.
Incorrect Answers
A:A token type must match the consuming API and the configured identity trust. It leaves the correct runtime principal and its bindings unaddressed.
B:Renaming a resource does not alter the actual effective configuration. The specific requirement concerns the correct runtime principal and its bindings. The appropriate action here is: Verify the instance's OAuth access scopes alongside IAM permissions and use compatible approved scopes for the workload.
D:Changing audit collection cannot grant a denied permission or fix service behavior. The necessary distinction is the correct runtime principal and its bindings.
Question 723
A team turns off every service account in a project to reduce security exposure and causes managed Google Cloud services to fail. What kind of identities must be distinguished?
- Inspect the service account's managed key metadata and available Cloud Audit Logs for service-account key creation and deletion events.
- Grant a new billing role instead of adjusting the actual token or impersonation permission.
- Identify Google-managed service agents and service-specific dependencies before modifying or disabling principals.
- Grant Organization Owner to the caller rather than validating the intended identity binding.
Explanation
Correct Answer
C:Some Google Cloud services use managed service agents to operate resources; disabling them blindly can break service functions.
Incorrect Answers
A:The suggestion does not establish the required issuer, principal or permission boundary. The specific requirement concerns the correct runtime principal and its bindings. The appropriate action here is: Identify Google-managed service agents and service-specific dependencies before modifying or disabling principals.
B:The suggestion does not establish the required issuer, principal or permission boundary. The necessary distinction is the correct runtime principal and its bindings. The appropriate action here is: Identify Google-managed service agents and service-specific dependencies before modifying or disabling principals.
D:Organization Owner is excessive privilege and cannot repair an unrelated configuration fault. The case instead depends on the correct runtime principal and its bindings.
Question 724
A platform engineer attempts to manually edit the key material of a Google-managed service agent as though it were a user-managed account. What is the correct approach?
- Inventory user-managed keys, establish rotation deadlines and migrate to keyless authentication where supported.
- Use the documented service-agent IAM and product configuration controls rather than directly managing Google-controlled keys.
- Rename the service account but leave the old principal and role bindings unchanged.
- Grant Organization Owner to the caller rather than validating the intended identity binding.
Explanation
Correct Answer
B:Service-agent credential lifecycle is handled by the managed service and differs from a customer-created service-account key.
Incorrect Answers
A:The suggestion does not establish the required issuer, principal or permission boundary. The necessary distinction is the workload's intended identity and permissions. The appropriate action here is: Use the documented service-agent IAM and product configuration controls rather than directly managing Google-controlled keys.
C:Renaming a resource does not alter the actual effective configuration. The case instead depends on the workload's intended identity and permissions.
D:Organization Owner is excessive privilege and cannot repair an unrelated configuration fault. It leaves the workload's intended identity and permissions unaddressed. The appropriate action here is: Use the documented service-agent IAM and product configuration controls rather than directly managing Google-controlled keys.
Question 725
A service-account credentials report shows a key is disabled, but a deployment owner asks to keep using it to avoid a release. What must happen?
- Migrate the deployment to a valid approved managed credential or reauthorized key through the documented security process.
- Update the staging deployment's credential reference to the approved new identity or key and redeploy it securely.
- Grant a new billing role instead of adjusting the actual token or impersonation permission.
- Reuse the default Compute Engine service account for every unrelated workload.
Explanation
Correct Answer
A:Disabled credentials cannot be relied upon for new authentication; production convenience does not override credential validity.
Incorrect Answers
B:Deploying a separate workload does not automatically configure the existing feature. The case instead depends on the workload's intended identity and permissions.
C:The suggestion does not establish the required issuer, principal or permission boundary. It leaves the workload's intended identity and permissions unaddressed. The appropriate action here is: Migrate the deployment to a valid approved managed credential or reauthorized key through the documented security process.
D:The suggestion does not establish the required issuer, principal or permission boundary. The specific requirement concerns the workload's intended identity and permissions. The appropriate action here is: Migrate the deployment to a valid approved managed credential or reauthorized key through the documented security process.
Question 726
A deployment operator has Compute Instance Admin in project blue but cannot launch a VM using the team-approved runtime service account. The IAM error cites iam.serviceAccounts.actAs on that service account. What additional permission is appropriate?
- Grant roles/iam.serviceAccountUser on that specific runtime service account to the authorized deployment operator.
- Grant Storage Object Viewer on an unrelated bucket, because VM creation implicitly consumes that permission.
- Grant the deployer an organization-wide Token Creator role on every service account as a substitute for actAs.
- Enable a larger VM CPU quota, since quota automatically provides missing IAM service-account attachment permissions.
Explanation
Correct Answer
A:Grant roles/iam.serviceAccountUser on that specific runtime service account to the authorized deployment operator. Service Account User contains iam.serviceAccounts.actAs and is required to attach the selected identity; the narrowly scoped grant completes the operator's existing VM creation permissions.
Incorrect Answers
B:Grant Storage Object Viewer on an unrelated bucket, because VM creation implicitly consumes that permission. Cloud Storage read access does not grant authority to attach a service account to a VM.
C:Grant the deployer an organization-wide Token Creator role on every service account as a substitute for actAs. Token Creator supports issuing short-lived credentials for impersonation and is broader/different from attaching an identity via actAs.
D:Enable a larger VM CPU quota, since quota automatically provides missing IAM service-account attachment permissions. Quota governs resource capacity, not IAM authorization to use a service account.
Question 727
A developer needs to obtain a short-lived access token as a test service account to debug one Cloud Storage request, but lacks token-generation permission. What role family is relevant?
- Have an authorized administrator enable the Service Account Credentials API in the relevant project, then verify token-creation permissions.
- Grant a new billing role instead of adjusting the actual token or impersonation permission.
- Grant Organization Owner to the caller rather than validating the intended identity binding.
- Grant the appropriate scoped Service Account Token Creator permission under a reviewed impersonation policy.
Explanation
Correct Answer
D:Generating an impersonated token is a different privilege from attaching a service account to a workload.
Incorrect Answers
A:The suggestion does not establish the required issuer, principal or permission boundary. The specific requirement concerns the permitted service-account delegation.
B:The suggestion does not establish the required issuer, principal or permission boundary. The necessary distinction is the permitted service-account delegation.
C:Organization Owner is excessive privilege and cannot repair an unrelated configuration fault. The case instead depends on the permitted service-account delegation.
Question 728
An engineer receives Service Account User on a production identity and assumes it automatically permits generating OAuth tokens from their laptop. What should be clarified?
- Remove the public delegation grant immediately and assess any minted credentials and audit activity under the account.
- Issue a long-lived service-account JSON key instead of repairing short-lived workload credentials.
- Require the separate token-minting permission where direct impersonated credentials are intended; do not confuse actAs with Token Creator.
- Rename the service account but leave the old principal and role bindings unchanged.
Explanation
Correct Answer
C:Service Account User and Token Creator authorize different operations and should be granted only when each use is justified.
Incorrect Answers
A:The suggestion does not establish the required issuer, principal or permission boundary. The necessary distinction is the permitted service-account delegation. The appropriate action here is: Require the separate token-minting permission where direct impersonated credentials are intended; do not confuse actAs with Token Creator.
B:The suggestion does not establish the required issuer, principal or permission boundary. The case instead depends on the permitted service-account delegation.
D:Renaming a resource does not alter the actual effective configuration. It leaves the permitted service-account delegation unaddressed.
Question 729
A CI pipeline uses gcloud with the personal developer identity even though a dedicated deployment service account exists. How should command execution be scoped?
- Correlate available STS and IAM Credentials audit events with service-account delegation information, mapped external subject attributes and target-resource logs.
- Issue a long-lived service-account JSON key instead of repairing short-lived workload credentials.
- Grant a new billing role instead of adjusting the actual token or impersonation permission.
- Use approved service-account impersonation for the deployment commands and record the originating caller in audit evidence.
Explanation
Correct Answer
D:Impersonation lets CI perform authorized actions under a managed workload principal without distributing private key material.
Incorrect Answers
A:The suggestion does not establish the required issuer, principal or permission boundary. The case instead depends on workload impersonation authorization.
B:The suggestion does not establish the required issuer, principal or permission boundary. It leaves workload impersonation authorization unaddressed.
C:The suggestion does not establish the required issuer, principal or permission boundary. The specific requirement concerns workload impersonation authorization.
Question 730
A release job can mint a token for a service account with broad production rights but the job only deploys to staging. What grant should be removed?
- Rotate the staging job credentials but leave its production Token Creator delegation intact.
- Remove production Token Creator delegation from the staging pipeline and use a separate staging deployment identity.
- Rename the service account but leave the old principal and role bindings unchanged.
- Issue a long-lived service-account JSON key instead of repairing short-lived workload credentials.
Explanation
Correct Answer
B:Service-account impersonation can indirectly grant all resource permissions of the target account; excessive delegation is high risk.
Incorrect Answers
A:Credential rotation does not revoke the production account impersonation grant available to staging.
C:Renaming a resource does not alter the actual effective configuration. The specific requirement concerns the permitted service-account delegation.
D:The suggestion does not establish the required issuer, principal or permission boundary. The necessary distinction is the permitted service-account delegation. The appropriate action here is: Remove production Token Creator delegation from the staging pipeline and use a separate staging deployment identity.