Google Professional Cloud Security Engineer Practice Test 75
Question 740
A deployer must perform an approved short-lived operation as a restricted deployment service account. After the recommended Service account impersonation and token creation change is implemented, which observation MOST directly verifies that it works?
- Approved requests yield bounded temporary privileges and unapproved users cannot self-elevate.
- Authorized pods read the assigned secret without node-wide credentials and other namespaces are denied.
- Audit evidence links the caller to short-lived impersonated credentials and unauthorized callers cannot impersonate.
- Developers can attach only designated execution identities and cannot elevate by choosing unrelated accounts.
Explanation
Correct Answer
C:Audit evidence links the caller to short-lived impersonated credentials and unauthorized callers cannot impersonate. This directly tests whether the Service account impersonation and token creation goal was met. Impersonation is an explicit IAM privilege and issues temporary credentials for another principal.
Incorrect Answers
A:Approved requests yield bounded temporary privileges and unapproved users cannot self-elevate. This is evidence about Privileged Access Manager just-in-time approval, not direct proof of the required Service account impersonation and token creation behavior. The relevant check is that audit evidence links the caller to short-lived impersonated credentials and unauthorized callers cannot impersonate.
B:Authorized pods read the assigned secret without node-wide credentials and other namespaces are denied. This is evidence about GKE workload identity binding, not direct proof of the required Service account impersonation and token creation behavior. The relevant check is that audit evidence links the caller to short-lived impersonated credentials and unauthorized callers cannot impersonate.
D:Developers can attach only designated execution identities and cannot elevate by choosing unrelated accounts. This is evidence about Scoped permission for service account creation, not direct proof of the required Service account impersonation and token creation behavior. The relevant check is that audit evidence links the caller to short-lived impersonated credentials and unauthorized callers cannot impersonate.
Question 741
A deployed AI assistant suddenly exhibits unusual tool invocation rates and retrieval requests. Which approach BEST meets the requirement while respecting the system constraints?
- Disable all model request monitoring so nobody can investigate unusual agent behavior.
- Monitor model and tool access logs, anomaly signals and permitted user identity context.
- Record trusted artifact provenance, integrity hashes and deployment approvals for model packages.
- Enforce tool authorization and distinguish untrusted retrieved content from trusted application instructions.
Explanation
Correct Answer
B:Monitor model and tool access logs, anomaly signals and permitted user identity context. AI runtime threat detection requires operational telemetry and governed diagnostic retention. Verification should demonstrate that incidents correlate sensitive operations to request identities without exposing raw private prompts unnecessarily.
Incorrect Answers
A:Disable all model request monitoring so nobody can investigate unusual agent behavior. This is an unsafe or ineffective shortcut for AI model monitoring and abuse investigations. No audit trail makes prompt injection or misuse difficult to reconstruct.
C:Record trusted artifact provenance, integrity hashes and deployment approvals for model packages. This primarily addresses AI model artifact provenance and integrity, not the specific AI model monitoring and abuse investigations decision. Here, aI runtime threat detection requires operational telemetry and governed diagnostic retention.
D:Enforce tool authorization and distinguish untrusted retrieved content from trusted application instructions. This primarily addresses Prompt injection and retrieved data boundary, not the specific AI model monitoring and abuse investigations decision. Here, aI runtime threat detection requires operational telemetry and governed diagnostic retention.
Question 742
A platform administrator should rotate keys but must not decrypt protected application data. After the recommended Cloud KMS crypto key role separation change is implemented, which observation MOST directly verifies that it works?
- Attestation and platform configuration show the expected confidential computing capability is enabled.
- Authorized applications connect privately while unrelated clients cannot reach or log into the database.
- Test users can administer key settings or decrypt only according to their distinct approved roles.
- Both sides authenticate the expected certificate chain and unauthorized service identities are denied.
Explanation
Correct Answer
C:Test users can administer key settings or decrypt only according to their distinct approved roles. This directly tests whether the Cloud KMS crypto key role separation goal was met. Cloud KMS separates administrative operations from cryptographic actions through IAM permissions.
Incorrect Answers
A: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 Cloud KMS crypto key role separation behavior. The relevant check is that test users can administer key settings or decrypt only according to their distinct approved roles.
B:Authorized applications connect privately while unrelated clients cannot reach or log into the database. This is evidence about Cloud SQL private IP and database authentication, not direct proof of the required Cloud KMS crypto key role separation behavior. The relevant check is that test users can administer key settings or decrypt only according to their distinct approved roles.
D:Both sides authenticate the expected certificate chain and unauthorized service identities are denied. This is evidence about Mutual TLS for service-to-service identity, not direct proof of the required Cloud KMS crypto key role separation behavior. The relevant check is that test users can administer key settings or decrypt only according to their distinct approved roles.
Question 743
A provider hosts a private TCP service that consumer VPC networks should access without VPC peering. After the recommended Private Service Connect service attachment publishing change is implemented, which observation MOST directly verifies that it works?
- Approved consumer endpoints connect privately while unauthorized projects cannot create effective connections.
- Unapproved API enablement and access attempts are prevented while needed services remain operational.
- Private workloads can reach supported Google APIs without an assigned external IP.
- On-premises clients can reach supported APIs through the intended private Google connectivity path.
Explanation
Correct Answer
A:Approved consumer endpoints connect privately while unauthorized projects cannot create effective connections. This directly tests whether the Private Service Connect service attachment publishing goal was met. PSC service attachments expose specific producer services rather than transitive access to the entire VPC.
Incorrect Answers
B:Unapproved API enablement and access attempts are prevented while needed services remain operational. This is evidence about API exposure restriction and service inventory, not direct proof of the required Private Service Connect service attachment publishing behavior. The relevant check is that approved consumer endpoints connect privately while unauthorized projects cannot create effective connections.
C:Private workloads can reach supported Google APIs without an assigned external IP. This is evidence about Private Google Access for internal VMs, not direct proof of the required Private Service Connect service attachment publishing behavior. The relevant check is that approved consumer endpoints connect privately while unauthorized projects cannot create effective connections.
D:On-premises clients can reach supported APIs through the intended private Google connectivity path. This is evidence about Private Google Access for on-premises, not direct proof of the required Private Service Connect service attachment publishing behavior. The relevant check is that approved consumer endpoints connect privately while unauthorized projects cannot create effective connections.
Question 744
A delivery service needs to issue access links that expire after a short contract window. An engineer must explain the recommendation for Time-bounded Cloud Storage signed URLs. Which underlying principle is the MOST accurate?
- Metadata server tokens reflect attached workload identity and must be protected from untrusted code paths.
- AI runtime threat detection requires operational telemetry and governed diagnostic retention.
- Signed URLs delegate bounded access through a signature rather than lasting IAM grants to each downloader.
- Uniform bucket-level access replaces object ACL authorization with IAM while public access prevention is separate.
Explanation
Correct Answer
C:Signed URLs delegate bounded access through a signature rather than lasting IAM grants to each downloader. Therefore the implementation decision is to issue a supported signed URL with limited method, object, expiry and protected signing identity.
Incorrect Answers
A:A widely forwarded valid signed URL can be used by others until expiry unless additional controls exist. Metadata server tokens reflect attached workload identity and must be protected from untrusted code paths.
B:A widely forwarded valid signed URL can be used by others until expiry unless additional controls exist. AI runtime threat detection requires operational telemetry and governed diagnostic retention.
D:A widely forwarded valid signed URL can be used by others until expiry unless additional controls exist. Uniform bucket-level access replaces object ACL authorization with IAM while public access prevention is separate.
Question 745
Regional managers share a BigQuery table but must see only their assigned geographic rows. After the recommended BigQuery row-level access policy change is implemented, which observation MOST directly verifies that it works?
- Test objects transition or expire according to policy while exempt legal artifacts are preserved.
- Output removes intended values and controlled inspection confirms no exposed matches in sample logs.
- The same query yields permitted rows for each test identity and denies or filters restricted rows.
- The workload retains required API access without broad metadata-derived permissions.
Explanation
Correct Answer
C:The same query yields permitted rows for each test identity and denies or filters restricted rows. This directly tests whether the BigQuery row-level access policy goal was met. Row access policies enforce table row restrictions independent of a user's optional SQL filter.
Incorrect Answers
A:Test objects transition or expire according to policy while exempt legal artifacts are preserved. This is evidence about Cloud Storage object lifecycle transition, not direct proof of the required BigQuery row-level access policy behavior. The relevant check is that the same query yields permitted rows for each test identity and denies or filters restricted rows.
B:Output removes intended values and controlled inspection confirms no exposed matches in sample logs. This is evidence about Sensitive Data Protection inspection redaction, not direct proof of the required BigQuery row-level access policy behavior. The relevant check is that the same query yields permitted rows for each test identity and denies or filters restricted rows.
D:The workload retains required API access without broad metadata-derived permissions. This is evidence about Compute metadata exposure hardening, not direct proof of the required BigQuery row-level access policy behavior. The relevant check is that the same query yields permitted rows for each test identity and denies or filters restricted rows.
Question 746
A deployment team needs application logs but should not read access logs containing sensitive request metadata. If the relevant Least privilege access to audit logs safeguard is omitted, which outcome is the MOST directly associated failure?
- Reviewing only workload stdout logs can miss changes to IAM or organization policies.
- An overprivileged builder can modify deployment images, secrets and policies beyond its intended pipeline.
- Giving all developers logging administrator and private-log access may expose sensitive audit metadata.
- Mirroring all production traffic to an untrusted collector creates a new exfiltration risk.
Explanation
Correct Answer
C:Giving all developers logging administrator and private-log access may expose sensitive audit metadata. Log viewer roles and sink or retention administration require distinct privileges. The corrective approach is to verify log view filters and permissions with multiple representative users.
Incorrect Answers
A:This failure pattern belongs chiefly to Admin Activity Audit Logs provenance; it is not the failure most directly tested in the Least privilege access to audit logs situation.
B:This failure pattern belongs chiefly to Secure Cloud Build execution identity; it is not the failure most directly tested in the Least privilege access to audit logs situation.
D:This failure pattern belongs chiefly to Packet Mirroring for intrusion investigation; it is not the failure most directly tested in the Least privilege access to audit logs situation.
Question 747
An organization needs an explicit prohibition for a high-risk operation across a folder. Which practical correction or implementation decision is MOST directly appropriate?
- Inspect policy inheritance and test both denied and exempt principals.
- Assume an allow binding overrides every matching higher-level IAM deny rule.
- Reproduce the authorization check with precise principal, permission and target resource.
- Use Policy Troubleshooter and negative tests to verify effective role scope.
Explanation
Correct Answer
A:Inspect policy inheritance and test both denied and exempt principals. The intended behavior can be checked because the prohibited operation is denied despite applicable allow bindings and legitimate exceptions remain functional.
Incorrect Answers
B:Assume an allow binding overrides every matching higher-level IAM deny rule. This would not correct the IAM Deny policy enforcement workflow. An overly broad deny rule can block critical service agents and break production operations.
C:Reproduce the authorization check with precise principal, permission and target resource. This could help IAM effective permission troubleshooting, but it does not resolve the IAM Deny policy enforcement case. IAM deny policy can restrict specified permissions notwithstanding an otherwise matching allow grant.
D:Use Policy Troubleshooter and negative tests to verify effective role scope. This could help IAM least-privilege role scoping, but it does not resolve the IAM Deny policy enforcement case. IAM deny policy can restrict specified permissions notwithstanding an otherwise matching allow grant.
Question 748
A partner's branch connects over VPN but should reach only one application subnet. Which approach BEST meets the requirement while respecting the system constraints?
- Put the proposed VPC Service Controls configuration in dry-run mode and inspect violations.
- Apply routed network segmentation and firewall policies to restrict hybrid source ranges and workload targets.
- Define least-privilege stateful VPC firewall ingress and egress controls for the intended connections.
- Assume encrypting the VPN tunnel automatically gives branch users project IAM permissions.
Explanation
Correct Answer
B:Apply routed network segmentation and firewall policies to restrict hybrid source ranges and workload targets. Hybrid connectivity creates a route path but application and subnet firewall policy still governs permitted flows. Verification should demonstrate that the connected branch accesses only the approved service and unrelated project traffic is denied.
Incorrect Answers
A:Put the proposed VPC Service Controls configuration in dry-run mode and inspect violations. This primarily addresses VPC Service Controls dry-run evaluation, not the specific Hybrid network segmentation policy placement decision. Here, hybrid connectivity creates a route path but application and subnet firewall policy still governs permitted flows.
C:Define least-privilege stateful VPC firewall ingress and egress controls for the intended connections. This primarily addresses VPC firewall stateful filtering, not the specific Hybrid network segmentation policy placement decision. Here, hybrid connectivity creates a route path but application and subnet firewall policy still governs permitted flows.
D:Assume encrypting the VPN tunnel automatically gives branch users project IAM permissions. This is an unsafe or ineffective shortcut for Hybrid network segmentation policy placement. Advertising every private prefix without policy limits can enable lateral access from on-premises.
Question 749
One VPN tunnel is down but regulated workloads require continued private connectivity. Which practical correction or implementation decision is MOST directly appropriate?
- Test tunnel failure and BGP route convergence without assuming a single active path is enough.
- Test API hostname resolution and reachability from a VM with no external IP.
- Map the source and destination traffic path before changing unrelated permission policies.
- Use a single static public TCP proxy and call it a redundant encrypted site-to-site VPN.
Explanation
Correct Answer
A:Test tunnel failure and BGP route convergence without assuming a single active path is enough. The intended behavior can be checked because traffic remains encrypted and routes converge when a single tunnel or interface fails.
Incorrect Answers
B:Test API hostname resolution and reachability from a VM with no external IP. This could help Private Google Access for internal VMs, but it does not resolve the HA VPN encrypted hybrid tunnels case. HA VPN provides IPsec-based encrypted connectivity while redundancy depends on correct tunnel and BGP design.
C:Map the source and destination traffic path before changing unrelated permission policies. This could help Private connectivity choice and API access troubleshooting, but it does not resolve the HA VPN encrypted hybrid tunnels case. HA VPN provides IPsec-based encrypted connectivity while redundancy depends on correct tunnel and BGP design.
D:Use a single static public TCP proxy and call it a redundant encrypted site-to-site VPN. This would not correct the HA VPN encrypted hybrid tunnels workflow. One misconfigured BGP peer can eliminate the claimed high-availability path.
Question 750
A payment backend sends personal data through a client connection that requires TLS validation. If the relevant TLS certificate trust and cipher policy safeguard is omitted, which outcome is the MOST directly associated failure?
- Removing the service agent's key permissions may make resource data unavailable.
- Broad delete rules can remove records required by retention policy.
- Skipping certificate verification permits active interception despite an encrypted connection.
- Ignoring legacy ACL consumers may break clients during migration.
Explanation
Correct Answer
C:Skipping certificate verification permits active interception despite an encrypted connection. Encryption in transit needs endpoint identity validation in addition to encrypted bytes. The corrective approach is to test certificate expiry, hostname checks and backend TLS under expected client behavior.
Incorrect Answers
A:Removing the service agent's key permissions may make resource data unavailable. This failure pattern belongs chiefly to Cloud KMS CMEK service agent access; it is not the failure most directly tested in the TLS certificate trust and cipher policy situation. Skipping certificate verification permits active interception despite an encrypted connection.
B:Broad delete rules can remove records required by retention policy. This failure pattern belongs chiefly to Cloud Storage object lifecycle transition; it is not the failure most directly tested in the TLS certificate trust and cipher policy situation. Skipping certificate verification permits active interception despite an encrypted connection.
D:Ignoring legacy ACL consumers may break clients during migration. This failure pattern belongs chiefly to Cloud Storage uniform bucket-level permissions; it is not the failure most directly tested in the TLS certificate trust and cipher policy situation. Skipping certificate verification permits active interception despite an encrypted connection.