Google Professional Cloud Security Engineer Practice Test 12
Question 110
An application team confuses enterprise employee SAML login with an OAuth authorization grant. After the recommended SAML versus OAuth authentication purpose change is implemented, which observation MOST directly verifies that it works?
- The scheduled synchronization creates and updates the intended users without unexpectedly deleting cloud identities.
- Old privileged sessions expire according to policy and authorized workers can reauthenticate.
- Controlled test logins work after the change and emergency recovery stays documented.
- Sign-in and API scopes are verified separately according to their respective trust boundaries.
Explanation
Correct Answer
D:Sign-in and API scopes are verified separately according to their respective trust boundaries. This directly tests whether the SAML versus OAuth authentication purpose goal was met. SAML assertions and OAuth access tokens have different audiences and security purposes.
Incorrect Answers
A:The scheduled synchronization creates and updates the intended users without unexpectedly deleting cloud identities. This is evidence about Cloud Identity directory synchronization, not direct proof of the required SAML versus OAuth authentication purpose behavior. The relevant check is that sign-in and API scopes are verified separately according to their respective trust boundaries.
B:Old privileged sessions expire according to policy and authorized workers can reauthenticate. This is evidence about Session duration and reauthentication policy, not direct proof of the required SAML versus OAuth authentication purpose behavior. The relevant check is that sign-in and API scopes are verified separately according to their respective trust boundaries.
C:Controlled test logins work after the change and emergency recovery stays documented. This is evidence about Identity-provider failover and recovery testing, not direct proof of the required SAML versus OAuth authentication purpose behavior. The relevant check is that sign-in and API scopes are verified separately according to their respective trust boundaries.
Question 111
A security auditor assumes Google Cloud patches every guest operating system in Compute Engine. A reviewer proposes this shortcut: Assume a cloud service provider automatically approves every customer's application IAM policy. Which principle BEST explains why a different solution is needed?
- PCI segmentation reduces scope and attack surface only when enforced and validated.
- Scoping is based on actual data and connected system boundaries rather than arbitrary project names.
- Kubernetes network policies can restrict pod traffic when enforced by the cluster's networking implementation.
- Shared responsibility varies across IaaS and managed cloud services and cannot be inferred from branding alone.
Explanation
Correct Answer
D:Shared responsibility varies across IaaS and managed cloud services and cannot be inferred from branding alone. The appropriate approach is to build a service-specific shared responsibility matrix allocating cloud provider and customer controls.
Incorrect Answers
A:PCI segmentation reduces scope and attack surface only when enforced and validated. This is about PCI workload network segmentation, not why the proposed Cloud shared responsibility service boundaries shortcut should be rejected. Omitting customer-owned guest patches or data authorization creates compliance gaps.
B:Scoping is based on actual data and connected system boundaries rather than arbitrary project names. This is about Regulatory scope inventory and boundaries, not why the proposed Cloud shared responsibility service boundaries shortcut should be rejected. Omitting customer-owned guest patches or data authorization creates compliance gaps.
C:Kubernetes network policies can restrict pod traffic when enforced by the cluster's networking implementation. This is about GKE network policy for pod segmentation, not why the proposed Cloud shared responsibility service boundaries shortcut should be rejected. Omitting customer-owned guest patches or data authorization creates compliance gaps.
Question 112
A VM has both internal and external IP addresses and security asks which path can reach a public client. Which practical correction or implementation decision is MOST directly appropriate?
- Test queries from associated and nonassociated networks and pair with network controls.
- Inventory forwarding rules, NAT paths, routes and service endpoint exposure.
- Verify load balancer/backend exposure and test authorized versus unauthorized identities.
- Assume choosing a 10.0.0.0/8 address automatically removes every internet-facing load balancer.
Explanation
Correct Answer
B:Inventory forwarding rules, NAT paths, routes and service endpoint exposure. The intended behavior can be checked because only intended networks or authenticated clients reach the workload while public scans fail.
Incorrect Answers
A:Test queries from associated and nonassociated networks and pair with network controls. This could help Cloud DNS response policy filtering, but it does not resolve the Private versus public endpoint design case. Private addressing changes routability but does not independently confer application authorization.
C:Verify load balancer/backend exposure and test authorized versus unauthorized identities. This could help Identity-Aware Proxy application access, but it does not resolve the Private versus public endpoint design case. Private addressing changes routability but does not independently confer application authorization.
D:Assume choosing a 10.0.0.0/8 address automatically removes every internet-facing load balancer. This would not correct the Private versus public endpoint design workflow. Assigning an internal IP while retaining an exposed public load balancer can leave ingress available.
Question 113
An auditor needs to identify which human impersonated a privileged service account during maintenance. Which practical correction or implementation decision is MOST directly appropriate?
- Test deployments with permitted and forbidden target service account identities.
- Review runtime identity bindings and test legitimate and prohibited API calls.
- Download a reusable private key and share it on a public team wiki instead of impersonation.
- Review impersonation IAM policy and correlate caller and delegate audit activity.
Explanation
Correct Answer
D:Review impersonation IAM policy and correlate caller and delegate audit activity. The intended behavior can be checked because audit evidence links the caller to short-lived impersonated credentials and unauthorized callers cannot impersonate.
Incorrect Answers
A:Test deployments with permitted and forbidden target service account identities. This could help Scoped permission for service account creation, but it does not resolve the Service account impersonation and token creation case. Impersonation is an explicit IAM privilege and issues temporary credentials for another principal.
B:Review runtime identity bindings and test legitimate and prohibited API calls. This could help Service account per-workload identity design, but it does not resolve the Service account impersonation and token creation case. Impersonation is an explicit IAM privilege and issues temporary credentials for another principal.
C:Download a reusable private key and share it on a public team wiki instead of impersonation. This would not correct the Service account impersonation and token creation workflow. Overbroad Token Creator grants allow unintended principals to act as the target service account.
Question 114
A healthcare analytics team uses Cloud Storage and BigQuery for patient-associated datasets. After the recommended Healthcare data privacy segmentation change is implemented, which observation MOST directly verifies that it works?
- Authorized investigators retrieve allowed aggregate values while raw patient fields remain denied.
- A sample alert is triaged, escalated and closed with attributable decision and evidence.
- Applicable requests require approved decisions and corresponding events remain traceable.
- Attestation and platform configuration show the expected confidential computing capability is enabled.
Explanation
Correct Answer
A:Authorized investigators retrieve allowed aggregate values while raw patient fields remain denied. This directly tests whether the Healthcare data privacy segmentation goal was met. Healthcare privacy controls require data minimization, authorization and auditability regardless of infrastructure encryption.
Incorrect Answers
B:A sample alert is triaged, escalated and closed with attributable decision and evidence. This is evidence about SOC control evidence and incident workflow, not direct proof of the required Healthcare data privacy segmentation behavior. The relevant check is that authorized investigators retrieve allowed aggregate values while raw patient fields remain denied.
C:Applicable requests require approved decisions and corresponding events remain traceable. This is evidence about Access Approval Google support access, not direct proof of the required Healthcare data privacy segmentation behavior. The relevant check is that authorized investigators retrieve allowed aggregate values while raw patient fields remain denied.
D:Attestation and platform configuration show the expected confidential computing capability is enabled. This is evidence about Confidential VM memory protection, not direct proof of the required Healthcare data privacy segmentation behavior. The relevant check is that authorized investigators retrieve allowed aggregate values while raw patient fields remain denied.
Question 115
A protected VPC Service Controls perimeter should limit network access to supported restricted Google APIs. After the recommended Restricted Google APIs VIP for perimeters change is implemented, which observation MOST directly verifies that it works?
- Test host reachability and exported routes without assuming implicit transit through another peer.
- Traffic remains encrypted and routes converge when a single tunnel or interface fails.
- Expected network prefixes appear in learned and advertised route tables after convergence.
- DNS lookup and API traffic use the restricted VIP while unsupported services are handled intentionally.
Explanation
Correct Answer
D:DNS lookup and API traffic use the restricted VIP while unsupported services are handled intentionally. This directly tests whether the Restricted Google APIs VIP for perimeters goal was met. Restricted Google APIs endpoints narrow accessible Google API services when correctly routed.
Incorrect Answers
A:Test host reachability and exported routes without assuming implicit transit through another peer. This is evidence about VPC Network Peering transitivity limits, not direct proof of the required Restricted Google APIs VIP for perimeters behavior. The relevant check is that dNS lookup and API traffic use the restricted VIP while unsupported services are handled intentionally.
B:Traffic remains encrypted and routes converge when a single tunnel or interface fails. This is evidence about HA VPN encrypted hybrid tunnels, not direct proof of the required Restricted Google APIs VIP for perimeters behavior. The relevant check is that dNS lookup and API traffic use the restricted VIP while unsupported services are handled intentionally.
C:Expected network prefixes appear in learned and advertised route tables after convergence. This is evidence about Cloud Router dynamic BGP advertisement, not direct proof of the required Restricted Google APIs VIP for perimeters behavior. The relevant check is that dNS lookup and API traffic use the restricted VIP while unsupported services are handled intentionally.
Question 116
A platform administrator should rotate keys but must not decrypt protected application data. Which approach BEST meets the requirement while respecting the system constraints?
- Distinguish Google-managed default at-rest encryption from services configured with supported CMEK.
- Separate key administration permissions from cryptographic use and decryption permissions.
- Give every KMS administrator an organization Owner role for convenience.
- Select a supported Cloud HSM protection level for the relevant Cloud KMS key.
Explanation
Correct Answer
B:Separate key administration permissions from cryptographic use and decryption permissions. Cloud KMS separates administrative operations from cryptographic actions through IAM permissions. Verification should demonstrate that test users can administer key settings or decrypt only according to their distinct approved roles.
Incorrect Answers
A:Distinguish Google-managed default at-rest encryption from services configured with supported CMEK. This primarily addresses Default at-rest encryption versus CMEK, not the specific Cloud KMS crypto key role separation decision. Here, cloud KMS separates administrative operations from cryptographic actions through IAM permissions.
C:Give every KMS administrator an organization Owner role for convenience. This is an unsafe or ineffective shortcut for Cloud KMS crypto key role separation. Combining key admin and decrypter grants broadly defeats separation of duties.
D:Select a supported Cloud HSM protection level for the relevant Cloud KMS key. This primarily addresses Cloud HSM versus software-protected keys, not the specific Cloud KMS crypto key role separation decision. Here, cloud KMS separates administrative operations from cryptographic actions through IAM permissions.
Question 117
An operations team must revoke one leaked secret version without revealing other production secrets. An engineer must explain the recommendation for Secret Manager versioning and least privilege. Which underlying principle is the MOST accurate?
- Column-level policy tags and data masking govern sensitive columns separately from general dataset access.
- Imported key material requires compatible algorithms and a protected import workflow.
- Secrets are governed objects with IAM and version lifecycle; runtime variables alone are not a secret vault.
- Uniform bucket-level access replaces object ACL authorization with IAM while public access prevention is separate.
Explanation
Correct Answer
C:Secrets are governed objects with IAM and version lifecycle; runtime variables alone are not a secret vault. Therefore the implementation decision is to store approved secrets in Secret Manager versions and grant secret access only to intended workloads.
Incorrect Answers
A:Column-level policy tags and data masking govern sensitive columns separately from general dataset access. This describes the separate BigQuery column-level policy tags and masking mechanism; the present case instead requires Secret Manager versioning and least privilege. Shared plaintext tokens in source repositories remain exposed after workload deployment.
B:Imported key material requires compatible algorithms and a protected import workflow. This describes the separate Cloud KMS imported key material mechanism; the present case instead requires Secret Manager versioning and least privilege. Shared plaintext tokens in source repositories remain exposed after workload deployment.
D:Uniform bucket-level access replaces object ACL authorization with IAM while public access prevention is separate. This describes the separate Cloud Storage uniform bucket-level permissions mechanism; the present case instead requires Secret Manager versioning and least privilege. Shared plaintext tokens in source repositories remain exposed after workload deployment.
Question 118
An operations team must revoke one leaked secret version without revealing other production secrets. Which practical correction or implementation decision is MOST directly appropriate?
- Test repeatability, key custody and reversal permissions with sample sensitive values.
- Store production API credentials in a public Dockerfile ARG and assume Secret Manager manages them.
- Verify secret-access audit events, version disablement and consumer refresh.
- Inventory object ACL behavior, test IAM transitions and apply public access prevention where required.
Explanation
Correct Answer
C:Verify secret-access audit events, version disablement and consumer refresh. The intended behavior can be checked because authorized workloads retrieve the needed version and revoked versions cannot be accessed.
Incorrect Answers
A:Test repeatability, key custody and reversal permissions with sample sensitive values. This could help Sensitive Data Protection deterministic de-identification, but it does not resolve the Secret Manager versioning and least privilege case. Secrets are governed objects with IAM and version lifecycle; runtime variables alone are not a secret vault.
B:Store production API credentials in a public Dockerfile ARG and assume Secret Manager manages them. This would not correct the Secret Manager versioning and least privilege workflow. Shared plaintext tokens in source repositories remain exposed after workload deployment.
D:This could help Cloud Storage uniform bucket-level permissions, but it does not resolve the Secret Manager versioning and least privilege case.
Question 119
A model training team wants to fine-tune using proprietary support logs. If the relevant AI training data isolation and least privilege safeguard is omitted, which outcome is the MOST directly associated failure?
- Default or shared broad access can disclose unrelated data during experimentation.
- Poor tweak or key management can weaken confidentiality and cause inconsistent decryptability.
- Hiding PII only in the browser DOM does not prevent it from being returned by the API.
- Overprivileged training pipelines can leak data into artifacts and logs.
Explanation
Correct Answer
D:Overprivileged training pipelines can leak data into artifacts and logs. AI training workloads inherit ordinary cloud identity and data governance responsibilities. The corrective approach is to test data access, job execution identity and output artifact permissions before training.
Incorrect Answers
A:This failure pattern belongs chiefly to Vertex AI training and pipeline service identities; it is not the failure most directly tested in the AI training data isolation and least privilege situation.
B:Poor tweak or key management can weaken confidentiality and cause inconsistent decryptability. This failure pattern belongs chiefly to Format-preserving encryption data handling; it is not the failure most directly tested in the AI training data isolation and least privilege situation. Overprivileged training pipelines can leak data into artifacts and logs.
C:This failure pattern belongs chiefly to AI output data loss prevention; it is not the failure most directly tested in the AI training data isolation and least privilege situation.
Question 120
A data owner discovers an object ACL exposing one file more widely than the bucket IAM policy. Which practical correction or implementation decision is MOST directly appropriate?
- Assume turning on uniform access automatically changes all encrypted object contents.
- Inventory object ACL behavior, test IAM transitions and apply public access prevention where required.
- Verify secret-access audit events, version disablement and consumer refresh.
- Validate expiry, method scope and distribution restrictions with test downloaders.
Explanation
Correct Answer
B:Inventory object ACL behavior, test IAM transitions and apply public access prevention where required. The intended behavior can be checked because authorized bucket principals retain access and object-specific ACL exceptions stop applying.
Incorrect Answers
A:Assume turning on uniform access automatically changes all encrypted object contents. This would not correct the Cloud Storage uniform bucket-level permissions workflow. Ignoring legacy ACL consumers may break clients during migration.
C:Verify secret-access audit events, version disablement and consumer refresh. This could help Secret Manager versioning and least privilege, but it does not resolve the Cloud Storage uniform bucket-level permissions case. Uniform bucket-level access replaces object ACL authorization with IAM while public access prevention is separate.
D:Validate expiry, method scope and distribution restrictions with test downloaders. This could help Time-bounded Cloud Storage signed URLs, but it does not resolve the Cloud Storage uniform bucket-level permissions case. Uniform bucket-level access replaces object ACL authorization with IAM while public access prevention is separate.