Google Associate Cloud Engineer Practice Test 38
Question 370
A Compute Engine VM with an eligible regional external IPv4 address needs Standard Network Service Tier despite the project default being Premium. Which configuration takes precedence when supported?
- Rename the VPC network Standard; a network display name changes all external routes to the Standard service tier.
- Set the eligible VM/external-IP resource's network tier explicitly to Standard and verify that its frontend mode supports it.
- Turn off Cloud Logging on the project; logging state selects the default network tier for instance internet paths.
- Move the VM into a multi-region Cloud Storage bucket; object bucket location determines the VM's external network tier.
Explanation
Correct Answer
B:Set the eligible VM/external-IP resource's network tier explicitly to Standard and verify that its frontend mode supports it. For supported Compute Engine external endpoints, a resource-level tier choice can override the project's default tier when creating or configuring the eligible address.
Incorrect Answers
A:Rename the VPC network Standard; a network display name changes all external routes to the Standard service tier. VPC names are resource labels and do not determine Premium-versus-Standard eligible external IP configuration.
C:Turn off Cloud Logging on the project; logging state selects the default network tier for instance internet paths. Cloud Logging collection is independent of Network Service Tier. Disabling logs does not request Standard IP routing.
D:Move the VM into a multi-region Cloud Storage bucket; object bucket location determines the VM's external network tier. Compute Engine VMs are not members of Cloud Storage buckets, and storage location does not configure their regional external IP tier.
Question 371
A company plans to switch from Standard to Premium automatically only during an existing Standard Tier outage. What limitation should it review?
- Set the eligible VM/external-IP resource's network tier explicitly to Standard and verify that its frontend mode supports it.
- Change the VPC display name and assume routes and firewall rules are automatically corrected.
- Plan capacity and tier selection before the outage; emergency tier switching isn't a guaranteed immediate failover path.
- Change only the DNS record even though the described control is a network policy or route.
Explanation
Correct Answer
C:Network Service Tier migration during a Standard outage is subject to constraints and is not a dependable automatic backup.
Incorrect Answers
A:Network settings and service or application controls must be distinguished. The specific requirement concerns the chosen Network Service Tier.
B:Display names and labels identify resources; they do not enforce their runtime behavior. The necessary distinction is the chosen Network Service Tier.
D:Network settings and service or application controls must be distinguished. The case instead depends on the chosen Network Service Tier.
Question 372
A security architect says Standard Tier weakens firewall IAM policy enforcement because it uses public ISP paths. Which separation matters?
- Use Standard Network Service Tier where supported after validating performance and feature requirements.
- Keep IAM and VPC firewall protections independently configured; choose the tier for network path, cost and latency.
- Create an unrelated subnet without changing the active routes or selected targets.
- Turn off firewall logging before identifying the effective rule handling the packets.
Explanation
Correct Answer
B:Network Service Tiers affect internet routing and cost, not the semantic meaning of IAM bindings or firewall allow rules.
Incorrect Answers
A:Network settings and service or application controls must be distinguished. The necessary distinction is the chosen Network Service Tier. The appropriate action here is: Keep IAM and VPC firewall protections independently configured; choose the tier for network path, cost and latency.
C:This does not establish the network route or policy needed by the traffic. The case instead depends on the chosen Network Service Tier.
D:This does not establish the network route or policy needed by the traffic. It leaves the chosen Network Service Tier unaddressed. The appropriate action here is: Keep IAM and VPC firewall protections independently configured; choose the tier for network path, cost and latency.
Question 373
A Shared VPC service project already has roles/compute.networkUser on host subnet regulated-a, but its VM creation is still denied. The effective organization policy specifically restricts which Shared VPC subnets service projects may use. Which setting should an administrator investigate?
- The inherited constraints/compute.restrictSharedVpcSubnetworks policy and whether regulated-a is on the permitted subnet list.
- A new billing account for the service project; payment automatically overrides a resource-location or subnet restriction.
- The instance's CPU utilization alerts; Monitoring automatically authorizes a subnet once utilization falls.
- A second duplicate Network User grant; repeated allows override organization policy constraints.
Explanation
Correct Answer
A:The inherited constraints/compute.restrictSharedVpcSubnetworks policy and whether regulated-a is on the permitted subnet list. Google's Shared VPC documentation identifies organization policy restrictions as separate from Network User role bindings. An effective subnet restriction can block an otherwise authorized attachment.
Incorrect Answers
B:A new billing account for the service project; payment automatically overrides a resource-location or subnet restriction. Cloud Billing links do not supersede organization policy that restricts which shared network resource can be used.
C:The instance's CPU utilization alerts; Monitoring automatically authorizes a subnet once utilization falls. Monitoring telemetry does not alter effective organization policy constraints for VM networking.
D:A second duplicate Network User grant; repeated allows override organization policy constraints. An allow role cannot negate a resource organization policy prohibition; duplicate grants leave the same restriction in force.
Question 374
A regional internal load balancer has working application processes, but health checks are blocked by a restrictive backend firewall policy. Operations must restore probe access without opening client ingress to the internet. What should be allowed?
- The documented load-balancer health-check source ranges or identities and the correct probe ports to the eligible backend targets.
- 0.0.0.0/0 to every backend port because all internal load balancers require unrestricted public ingress for health checks.
- Only Cloud Billing API addresses; billing-account identity is how health probes reach internal backends.
- A route from Cloud SQL to the backend, because database routing replaces load-balancer health-check packets.
Explanation
Correct Answer
A:The documented load-balancer health-check source ranges or identities and the correct probe ports to the eligible backend targets. A narrowly targeted allow for Google's relevant health probe sources and ports lets the health check reach the serving process while respecting backend firewall boundaries.
Incorrect Answers
B:0.0.0.0/0 to every backend port because all internal load balancers require unrestricted public ingress for health checks. Allowing the entire public internet on every port is far broader than the probe requirement and creates an unnecessary attack surface.
C:Only Cloud Billing API addresses; billing-account identity is how health probes reach internal backends. Billing APIs report or collect charges; their endpoints are not the health-check source that probes application backends.
D:A route from Cloud SQL to the backend, because database routing replaces load-balancer health-check packets. Database routing does not authorize the load balancer's health-check packets through a backend VPC firewall.
Question 375
Two HA VPN tunnels show ESTABLISHED in the tunnel status, but on-premises hosts still cannot reach a newly created Google Cloud subnet because BGP exchanges omit that prefix. Which fix targets routing?
- Check the Cloud Router BGP advertised/learned prefixes and the on-premises router's accepted routes, then add the approved subnet advertisement.
- Delete and recreate the encrypted tunnels even though both are established; BGP advertisements are generated only by new tunnel keys.
- Grant VM Instance Admin to the user; VM administration causes the peer router to learn new subnet prefixes.
- Disable the subnet's IAM bindings; IAM allow bindings are the dynamic route advertisements for HA VPN peers.
Explanation
Correct Answer
A:Check the Cloud Router BGP advertised/learned prefixes and the on-premises router's accepted routes, then add the approved subnet advertisement. Tunnel availability and BGP route presence are separate states. The missing advertised prefix must be exchanged and accepted for packets to reach that subnet.
Incorrect Answers
B:Delete and recreate the encrypted tunnels even though both are established; BGP advertisements are generated only by new tunnel keys. Tunnel recreation risks disruption and does not automatically correct custom advertised-route settings on the Cloud Router or remote peer.
C:Grant VM Instance Admin to the user; VM administration causes the peer router to learn new subnet prefixes. Compute Engine IAM controls who may manage VMs; it does not program on-premises BGP routes.
D:Disable the subnet's IAM bindings; IAM allow bindings are the dynamic route advertisements for HA VPN peers. IAM principals and role grants are unrelated to Cloud Router's BGP route advertisements for private network prefixes.
Question 376
A Dedicated Interconnect circuit is operational and a VLAN attachment exists, but its attached VPC network has no dynamic routes to the customer network. Which routing configuration must be established?
- Associate the eligible VLAN attachment with a Cloud Router interface and configure a working BGP peer session with approved prefixes.
- Increase Cloud Storage replication; bucket redundancy automatically advertises on-premises CIDR routes over Interconnect.
- Enable Cloud CDN on the VPC; CDN serves as the BGP router for private VLAN attachments.
- Add an external IP to every VM; VLAN attachments exchange routes only when the VMs are reachable over public internet.
Explanation
Correct Answer
A:Associate the eligible VLAN attachment with a Cloud Router interface and configure a working BGP peer session with approved prefixes. A VLAN attachment and Cloud Router BGP session exchange dynamic routes with the on-premises router. A live physical circuit alone does not populate those VPC routes.
Incorrect Answers
B:Increase Cloud Storage replication; bucket redundancy automatically advertises on-premises CIDR routes over Interconnect. Object storage replication cannot configure VLAN attachment BGP interfaces or tell the customer router which VPC subnets to reach.
C:Enable Cloud CDN on the VPC; CDN serves as the BGP router for private VLAN attachments. Cloud CDN caches HTTP content and has no function as the dynamic routing peer for Dedicated Interconnect.
D:Add an external IP to every VM; VLAN attachments exchange routes only when the VMs are reachable over public internet. Interconnect is used for private network connectivity, and its route exchange does not require exposing internal VMs on public addresses.
Question 377
A global external Application Load Balancer currently has a healthy backend only in region A. The application is also deployed in region B, but that deployment is not attached to its backend service. How should the global frontend survive failure of region A?
- Attach and verify a qualified healthy region-B backend in the supported global backend service, testing its cross-region routing and failure handling.
- Raise the region-A VM CPU quota; higher single-region capacity protects the service against total regional unavailability.
- Change the URL map hostname to region-B text; a hostname string moves backend traffic without a backend-service configuration.
- Add a Cloud Billing account in region B; payer redundancy creates global load-balancer backend replication.
Explanation
Correct Answer
A:Attach and verify a qualified healthy region-B backend in the supported global backend service, testing its cross-region routing and failure handling. Global load balancing can select healthy backends in multiple regions only when those backends are actually configured in its serving backend service.
Incorrect Answers
B:Raise the region-A VM CPU quota; higher single-region capacity protects the service against total regional unavailability. CPU quota concerns capacity allocation in a still-operating region. It does not bring a fully unavailable region back online.
C:Change the URL map hostname to region-B text; a hostname string moves backend traffic without a backend-service configuration. URL mapping chooses a configured service; an arbitrary region name in a hostname does not attach an absent regional backend.
D:Add a Cloud Billing account in region B; payer redundancy creates global load-balancer backend replication. Cloud Billing accounts do not provision load-balancer backend resources or create automatic cross-region failover targets.
Question 378
Security deploys a global external HTTPS Application Load Balancer with supported edge policies. Finance proposes moving its frontend to Standard Network Service Tier while keeping the same global design. Which conclusion follows Google's tier matrix?
- Every global external Application Load Balancer supports Standard Tier if Cloud Armor is attached.
- The Network Service Tier only affects internal east-west VM traffic, so a global internet-facing HTTPS frontend is unaffected.
- Keep Premium Tier for the global external load balancer; evaluate an eligible regional external design separately if Standard Tier is mandatory.
- Choose Standard Tier and disable TLS termination so the global anycast HTTPS frontend remains supported.
Explanation
Correct Answer
C:Keep Premium Tier for the global external load balancer; evaluate an eligible regional external design separately if Standard Tier is mandatory. Google specifies Premium Tier for the global mode. Standard is supported for certain regional external modes, which are different architectures.
Incorrect Answers
A:Every global external Application Load Balancer supports Standard Tier if Cloud Armor is attached. An edge security policy does not change the published tier compatibility of a global external Application Load Balancer.
B:The Network Service Tier only affects internal east-west VM traffic, so a global internet-facing HTTPS frontend is unaffected. Tier selection changes routing and supported modes for external internet traffic; it is not limited to east-west communications.
D:Choose Standard Tier and disable TLS termination so the global anycast HTTPS frontend remains supported. Disabling TLS would violate the HTTPS requirement and does not convert the unsupported global Standard-tier combination into a valid one.
Question 379
A Terraform plan in CI proposes recreating a production database because a resource was manually renamed in the Google Cloud console, even though the database remains healthy. What should the operator reconcile?
- Create and retain a reviewed Terraform plan artifact, bind its approval to the deployment, and apply that exact plan.
- Disable all drift detection rather than investigating the actual resource change.
- Inspect Terraform state and the actual Google Cloud resource identity, then use a reviewed state move/import strategy rather than applying a destructive replacement blind.
- Discard the approved plan artifact and deploy a newly generated plan without comparison.
Explanation
Correct Answer
C:Terraform diffs reflect state and configuration addresses; reconciling managed identity safely matters more than blindly accepting replacement actions.
Incorrect Answers
A:Deploying a separate workload does not automatically configure the existing feature. The specific requirement concerns the approved Terraform plan and target.
B:The operation does not guarantee the correct managed resource or approved plan. The necessary distinction is the approved Terraform plan and target.
D:The operation does not guarantee the correct managed resource or approved plan. The case instead depends on the approved Terraform plan and target.
Question 380
Two release pipelines apply Terraform to the same shared Google Cloud state simultaneously, causing one deployment to overwrite assumptions from the other. What control prevents this?
- Inspect Terraform state and the actual Google Cloud resource identity, then use a reviewed state move/import strategy rather than applying a destructive replacement blind.
- Edit the Terraform state JSON manually without reconciling it to the real managed resource.
- Duplicate production credentials into a staging repository to bypass workspace selection.
- Use an approved shared remote backend with state locking and serialize protected production applies.
Explanation
Correct Answer
D:Concurrent state mutation can corrupt the intended configuration view; proper locking and change coordination limit conflicts.
Incorrect Answers
A:The operation does not guarantee the correct managed resource or approved plan. The necessary distinction is the approved Terraform plan and target. The appropriate action here is: Use an approved shared remote backend with state locking and serialize protected production applies.
B:Editing state by hand can desynchronize Terraform from the real infrastructure. The case instead depends on the approved Terraform plan and target.
C:Bypassing the guardrail weakens security without correctly configuring the resource. It leaves the approved Terraform plan and target unaddressed.