Google Professional Cloud Security Engineer Practice Test 83
Question 820
An organization needs an explicit prohibition for a high-risk operation across a folder. Testing exposes this failure: An overly broad deny rule can block critical service agents and break production operations. Which remediation should be applied FIRST?
- Define and test an Access Context Manager access level with intended device and network conditions.
- Assume an allow binding overrides every matching higher-level IAM deny rule.
- Grant resource roles to centrally governed groups and separately control membership administration.
- Inspect policy inheritance and test both denied and exempt principals.
Explanation
Correct Answer
D:Inspect policy inheritance and test both denied and exempt principals. Confirmation requires seeing that the prohibited operation is denied despite applicable allow bindings and legitimate exceptions remain functional.
Incorrect Answers
A:Define and test an Access Context Manager access level with intended device and network conditions. It addresses Access Context Manager access levels, not the diagnosed IAM Deny policy enforcement failure. An overly broad deny rule can block critical service agents and break production operations.
B:Assume an allow binding overrides every matching higher-level IAM deny rule. This retains the IAM Deny policy enforcement problem. IAM deny policy can restrict specified permissions notwithstanding an otherwise matching allow grant.
C:Grant resource roles to centrally governed groups and separately control membership administration. It addresses Group-based IAM authorization, not the diagnosed IAM Deny policy enforcement failure. An overly broad deny rule can block critical service agents and break production operations.
Question 821
A document retrieval agent reads a user-uploaded file containing instructions to reveal other users' records. If the relevant Prompt injection and retrieved data boundary safeguard is omitted, which outcome is the MOST directly associated failure?
- Treating retrieved text as a privileged system command can disclose protected data.
- Overprivileged training pipelines can leak data into artifacts and logs.
- No audit trail makes prompt injection or misuse difficult to reconstruct.
- Combining key admin and decrypter grants broadly defeats separation of duties.
Explanation
Correct Answer
A:Treating retrieved text as a privileged system command can disclose protected data. Prompt-injection defenses require trusted instruction boundaries and downstream authorization, not only content filters. The corrective approach is to test malicious retrieval fixtures and verify independent permission checks on every sensitive tool.
Incorrect Answers
B:This failure pattern belongs chiefly to AI training data isolation and least privilege; it is not the failure most directly tested in the Prompt injection and retrieved data boundary situation.
C:This failure pattern belongs chiefly to AI model monitoring and abuse investigations; it is not the failure most directly tested in the Prompt injection and retrieved data boundary situation.
D:This failure pattern belongs chiefly to Cloud KMS crypto key role separation; it is not the failure most directly tested in the Prompt injection and retrieved data boundary situation.
Question 822
A regulated project requires data encrypted under a KMS key in the expected location. An engineer must explain the recommendation for Cloud KMS CMEK service agent access. Which underlying principle is the MOST accurate?
- Minimization and lifecycle governance reduce residual sensitive-data exposure beyond IAM encryption.
- Cloud EKM shifts external key custody and introduces dependency on external key-management availability.
- Locked retention policies can enforce immutability even against privileged operators.
- CMEK integrates service-side encryption with explicit key permissions and regional compatibility.
Explanation
Correct Answer
D:CMEK integrates service-side encryption with explicit key permissions and regional compatibility. Therefore the implementation decision is to configure an eligible Cloud KMS CMEK key and grant the product service agent required crypto permissions.
Incorrect Answers
A:Minimization and lifecycle governance reduce residual sensitive-data exposure beyond IAM encryption. This describes the separate Sensitive data retention and data lifecycle mechanism; the present case instead requires Cloud KMS CMEK service agent access. Removing the service agent's key permissions may make resource data unavailable.
B:Cloud EKM shifts external key custody and introduces dependency on external key-management availability. This describes the separate External Key Manager custody mechanism; the present case instead requires Cloud KMS CMEK service agent access. Removing the service agent's key permissions may make resource data unavailable.
C:Locked retention policies can enforce immutability even against privileged operators. This describes the separate Cloud Storage retention policy and Bucket Lock mechanism; the present case instead requires Cloud KMS CMEK service agent access. Removing the service agent's key permissions may make resource data unavailable.
Question 823
A compromised application process attempts to obtain credentials from the VM metadata service. If the relevant Compute metadata exposure hardening safeguard is omitted, which outcome is the MOST directly associated failure?
- Treating retrieved text as a privileged system command can disclose protected data.
- Sharing the underlying dataset viewer role can reveal more than the intended view.
- Scanning sensitive data into an openly shared findings table can create a second disclosure path.
- An instance with Owner-level credentials makes metadata token theft especially damaging.
Explanation
Correct Answer
D:An instance with Owner-level credentials makes metadata token theft especially damaging. Metadata server tokens reflect attached workload identity and must be protected from untrusted code paths. The corrective approach is to test workload permissions, request paths and software isolation under least privilege.
Incorrect Answers
A:Treating retrieved text as a privileged system command can disclose protected data. This failure pattern belongs chiefly to Prompt injection and retrieved data boundary; it is not the failure most directly tested in the Compute metadata exposure hardening situation. An instance with Owner-level credentials makes metadata token theft especially damaging.
B:Sharing the underlying dataset viewer role can reveal more than the intended view. This failure pattern belongs chiefly to BigQuery authorized views; it is not the failure most directly tested in the Compute metadata exposure hardening situation. An instance with Owner-level credentials makes metadata token theft especially damaging.
C:This failure pattern belongs chiefly to Sensitive Data Protection PII inspection; it is not the failure most directly tested in the Compute metadata exposure hardening situation.
Question 824
A contractor should access one HTTPS service only when identity and device conditions qualify. Testing exposes this failure: Exposing a direct backend path may bypass the intended identity gate. Which remediation should be applied FIRST?
- Put an IAM role in the web page HTML and assume the browser cannot access the app without it.
- Verify load balancer/backend exposure and test authorized versus unauthorized identities.
- Inventory enabled services and apply supported service-use IAM or organization restrictions.
- Configure supported Cloud Armor throttling or rate-based ban rules at suitable match scope.
Explanation
Correct Answer
B:Verify load balancer/backend exposure and test authorized versus unauthorized identities. Confirmation requires seeing that signed-in authorized users reach the app and rejected identities cannot bypass the IAP path.
Incorrect Answers
A:Put an IAM role in the web page HTML and assume the browser cannot access the app without it. This retains the Identity-Aware Proxy application access problem. IAP adds identity-based access enforcement before reaching supported applications.
C:Inventory enabled services and apply supported service-use IAM or organization restrictions. It addresses API exposure restriction and service inventory, not the diagnosed Identity-Aware Proxy application access failure. Exposing a direct backend path may bypass the intended identity gate.
D:Configure supported Cloud Armor throttling or rate-based ban rules at suitable match scope. It addresses Cloud Armor rate limiting and bot mitigation, not the diagnosed Identity-Aware Proxy application access failure. Exposing a direct backend path may bypass the intended identity gate.
Question 825
A security analyst receives an alert of possible leaked service account token usage. After the recommended Incident response alert routing and containment change is implemented, which observation MOST directly verifies that it works?
- Test logs from multiple child projects arrive at the designated restricted destination.
- The selected instances install approved updates and show patch compliance and failures.
- Unauthorized access stops and the investigation retains time-stamped evidence and affected-resource inventory.
- Expected source-destination flow metadata appears without claiming every packet payload is captured.
Explanation
Correct Answer
C:Unauthorized access stops and the investigation retains time-stamped evidence and affected-resource inventory. This directly tests whether the Incident response alert routing and containment goal was met. Containment, evidence preservation and recovery are coordinated security operations rather than one dashboard click.
Incorrect Answers
A:Test logs from multiple child projects arrive at the designated restricted destination. This is evidence about Organization aggregated log sink, not direct proof of the required Incident response alert routing and containment behavior. The relevant check is that unauthorized access stops and the investigation retains time-stamped evidence and affected-resource inventory.
B:The selected instances install approved updates and show patch compliance and failures. This is evidence about VM Manager patch orchestration, not direct proof of the required Incident response alert routing and containment behavior. The relevant check is that unauthorized access stops and the investigation retains time-stamped evidence and affected-resource inventory.
D:Expected source-destination flow metadata appears without claiming every packet payload is captured. This is evidence about VPC Flow Logs traffic analysis, not direct proof of the required Incident response alert routing and containment behavior. The relevant check is that unauthorized access stops and the investigation retains time-stamped evidence and affected-resource inventory.
Question 826
A provider hosts a private TCP service that consumer VPC networks should access without VPC peering. Which approach BEST meets the requirement while respecting the system constraints?
- Configure supported restricted.googleapis.com DNS and routing where the perimeter design requires it.
- Publish a supported PSC service attachment with consumer acceptance policy and internal load balancer.
- Treat PSC service attachment creation as a license to route every subnet in the provider VPC.
- Configure Cloud NAT with appropriate gateway, addresses and subnet egress coverage.
Explanation
Correct Answer
B:Publish a supported PSC service attachment with consumer acceptance policy and internal load balancer. PSC service attachments expose specific producer services rather than transitive access to the entire VPC. Verification should demonstrate that approved consumer endpoints connect privately while unauthorized projects cannot create effective connections.
Incorrect Answers
A:Configure supported restricted.googleapis.com DNS and routing where the perimeter design requires it. This primarily addresses Restricted Google APIs VIP for perimeters, not the specific Private Service Connect service attachment publishing decision. Here, pSC service attachments expose specific producer services rather than transitive access to the entire VPC.
C:Treat PSC service attachment creation as a license to route every subnet in the provider VPC. This is an unsafe or ineffective shortcut for Private Service Connect service attachment publishing. An all-consumer publish policy may expose internal service connectivity beyond the contract.
D:Configure Cloud NAT with appropriate gateway, addresses and subnet egress coverage. This primarily addresses Cloud NAT outbound without inbound exposure, not the specific Private Service Connect service attachment publishing decision. Here, pSC service attachments expose specific producer services rather than transitive access to the entire VPC.
Question 827
A security team needs to limit what privileges an exposed workload identity could use. Select TWO statements that correctly combine the Compute metadata exposure hardening method with evidence or its governing principle.
- Assume hiding the metadata IP address in application comments prevents compromised code from reaching it.
- The workload retains required API access without broad metadata-derived permissions.
- Issue a supported signed URL with limited method, object, expiry and protected signing identity.
- Use dedicated least-privileged runtime identities and harden metadata access according to platform support.
- Configure a supported BigQuery row access policy tied to approved identities and predicates.
Explanation
Correct Answers
B:The workload retains required API access without broad metadata-derived permissions. This is directly observable validation of the Compute metadata exposure hardening result.
D:Use dedicated least-privileged runtime identities and harden metadata access according to platform support. This is the intended action for Compute metadata exposure hardening.
Incorrect Answers
A:Assume hiding the metadata IP address in application comments prevents compromised code from reaching it. This is not part of a safe or effective Compute metadata exposure hardening design. An instance with Owner-level credentials makes metadata token theft especially damaging.
C:Issue a supported signed URL with limited method, object, expiry and protected signing identity. This serves the different Time-bounded Cloud Storage signed URLs objective, not the selected Compute metadata exposure hardening plan. Metadata server tokens reflect attached workload identity and must be protected from untrusted code paths.
E:Configure a supported BigQuery row access policy tied to approved identities and predicates. This serves the different BigQuery row-level access policy objective, not the selected Compute metadata exposure hardening plan. Metadata server tokens reflect attached workload identity and must be protected from untrusted code paths.
Question 828
A company forwards selected security events to its external SIEM over approved channels. After the recommended External SIEM log export privacy change is implemented, which observation MOST directly verifies that it works?
- Authorized test reads produce the expected data access evidence while logging costs are understood.
- Build logs show permitted operations and deny unexpected production resource access.
- Test logs from multiple child projects arrive at the designated restricted destination.
- The SIEM receives the intended event types while restricted fields and destinations comply with policy.
Explanation
Correct Answer
D:The SIEM receives the intended event types while restricted fields and destinations comply with policy. This directly tests whether the External SIEM log export privacy goal was met. Log export requires authorized destination write paths, filters and governance beyond local logging.
Incorrect Answers
A:Authorized test reads produce the expected data access evidence while logging costs are understood. This is evidence about Data Access Audit Logs enablement and cost, not direct proof of the required External SIEM log export privacy behavior. The relevant check is that the SIEM receives the intended event types while restricted fields and destinations comply with policy.
B:Build logs show permitted operations and deny unexpected production resource access. This is evidence about Secure Cloud Build execution identity, not direct proof of the required External SIEM log export privacy behavior. The relevant check is that the SIEM receives the intended event types while restricted fields and destinations comply with policy.
C:Test logs from multiple child projects arrive at the designated restricted destination. This is evidence about Organization aggregated log sink, not direct proof of the required External SIEM log export privacy behavior. The relevant check is that the SIEM receives the intended event types while restricted fields and destinations comply with policy.
Question 829
A company needs encrypted traffic between its data center and Google Cloud with redundant tunnel availability. After the recommended HA VPN encrypted hybrid tunnels change is implemented, which observation MOST directly verifies that it works?
- Logs identify would-be denied API calls while existing workloads continue to operate.
- DNS lookup and API traffic use the restricted VIP while unsupported services are handled intentionally.
- Approved consumer endpoints connect privately while unauthorized projects cannot create effective connections.
- Traffic remains encrypted and routes converge when a single tunnel or interface fails.
Explanation
Correct Answer
D:Traffic remains encrypted and routes converge when a single tunnel or interface fails. This directly tests whether the HA VPN encrypted hybrid tunnels goal was met. HA VPN provides IPsec-based encrypted connectivity while redundancy depends on correct tunnel and BGP design.
Incorrect Answers
A:Logs identify would-be denied API calls while existing workloads continue to operate. This is evidence about VPC Service Controls dry-run evaluation, not direct proof of the required HA VPN encrypted hybrid tunnels behavior. The relevant check is that traffic remains encrypted and routes converge when a single tunnel or interface fails.
B:DNS lookup and API traffic use the restricted VIP while unsupported services are handled intentionally. This is evidence about Restricted Google APIs VIP for perimeters, not direct proof of the required HA VPN encrypted hybrid tunnels behavior. The relevant check is that traffic remains encrypted and routes converge when a single tunnel or interface fails.
C:Approved consumer endpoints connect privately while unauthorized projects cannot create effective connections. This is evidence about Private Service Connect service attachment publishing, not direct proof of the required HA VPN encrypted hybrid tunnels behavior. The relevant check is that traffic remains encrypted and routes converge when a single tunnel or interface fails.
Question 830
An internal legacy report expects a fixed-format identifier but must not see a clear customer number. If the relevant Format-preserving encryption data handling safeguard is omitted, which outcome is the MOST directly associated failure?
- Sharing the underlying dataset viewer role can reveal more than the intended view.
- Shared plaintext tokens in source repositories remain exposed after workload deployment.
- Trusting all clients solely because they are inside the VPC may allow lateral impersonation.
- Poor tweak or key management can weaken confidentiality and cause inconsistent decryptability.
Explanation
Correct Answer
D:Poor tweak or key management can weaken confidentiality and cause inconsistent decryptability. Format-preserving encryption retains structural constraints but is not the same as plain-text masking. The corrective approach is to validate format rules, deterministic behavior where required and cryptographic key permissions.
Incorrect Answers
A:Sharing the underlying dataset viewer role can reveal more than the intended view. This failure pattern belongs chiefly to BigQuery authorized views; it is not the failure most directly tested in the Format-preserving encryption data handling situation. Poor tweak or key management can weaken confidentiality and cause inconsistent decryptability.
B:Shared plaintext tokens in source repositories remain exposed after workload deployment. This failure pattern belongs chiefly to Secret Manager versioning and least privilege; it is not the failure most directly tested in the Format-preserving encryption data handling situation. Poor tweak or key management can weaken confidentiality and cause inconsistent decryptability.
C:This failure pattern belongs chiefly to Mutual TLS for service-to-service identity; it is not the failure most directly tested in the Format-preserving encryption data handling situation.