Google Associate Cloud Engineer Practice Test 12
Question 110
An organization-wide Cloud Asset Inventory search returns zero results for a newly deployed product type. The engineer suspects this asset type is unsupported by the searchResources API. What is the correct first check?
- Verify that this exact asset type is supported by the selected Cloud Asset Inventory search method; use its product API if it is not.
- Remove all asset-type filters and conclude the resource does not exist if a second search is empty.
- Increase the IAM role to organization Owner because every Google resource is guaranteed to be indexed by every asset search interface.
- Change the Cloud Storage bucket location because asset search only inventories resources in the bucket's region.
Explanation
Correct Answer
A:Verify that this exact asset type is supported by the selected Cloud Asset Inventory search method; use its product API if it is not. Cloud Asset Inventory supports specific asset types and content in each API. An unsupported type cannot be discovered through that search method regardless of resource existence.
Incorrect Answers
B:Remove all asset-type filters and conclude the resource does not exist if a second search is empty. Clearing a filter does not change the API's catalog of supported asset types. An empty result is not proof that a product API has no resources.
C:Increase the IAM role to organization Owner because every Google resource is guaranteed to be indexed by every asset search interface. Broader permissions cannot add an unsupported asset type to an API. The search feature's coverage must be checked before interpreting missing results.
D:Change the Cloud Storage bucket location because asset search only inventories resources in the bucket's region. Inventory search operates over resource scope and indexed types, not the location of an unrelated Cloud Storage bucket. That change will not add indexing coverage.
Question 111
A team needs an inventory with the exact state of all managed instances at a specified compliance cutoff timestamp. Which method handles the historical requirement?
- Search the current Resource Manager project list and label the results with last month’s date.
- Use a Cloud Asset Inventory historical snapshot/export at the requested read time, if it falls within supported history.
- Collect the current Monitoring CPU metrics and assume they describe last month’s resource configuration.
- Rename present-day VM instances to match their old audit names.
Explanation
Correct Answer
B:Asset Inventory supports time-bounded asset history where the requested read time is within its retention window. Use a point-in-time snapshot or retained export with its timestamp recorded; for older cutoffs, rely on archived evidence captured then.
Incorrect Answers
A:A current list cannot prove that resources had the same configuration at the earlier compliance cutoff.
C:Performance metrics describe usage but not the complete historical resource inventory at a specified instant.
D:Renaming present resources cannot reconstruct deleted or changed instances from the earlier date.
Question 112
An engineer wants to find all resources in a folder but passes the folder's display name as a project path in an asset search. How should the scope be specified?
- Use the folder display name as if it were a project ID in every asset search.
- Replace the scope with a billing-account name since billing accounts own folders.
- Query a random project and assume its results include all sibling folder descendants.
- Use the resource scope in the correct form, such as folders/FOLDER_NUMBER for the target folder.
Explanation
Correct Answer
D:Cloud Asset Inventory searches use explicit scope resource paths. Obtain the target folder’s actual numeric identifier and pass the documented folder scope so the search covers the intended hierarchy rather than a named project.
Incorrect Answers
A:Display names are not canonical API resource identifiers and may be duplicated or changed.
B:Billing accounts do not define folders or replace their Resource Manager identifiers.
C:A project-scoped query would omit other projects under the intended folder.
Question 113
Gemini Cloud Assist recommends opening a firewall ingress rule to 0.0.0.0/0 to fix an internal service timeout, but the affected resources have not been inspected. What should the operator do before changing production networking?
- Apply the generated rule directly because assistant suggestions always override network security requirements.
- Give every workload account organization Owner so the assistant no longer needs to inspect firewall policies.
- Disable Firewall Rules Logging, which validates that the generated ingress change is safe for every target.
- Inspect the actual packet path, firewall policy and service identity; review a least-privilege change before deployment.
Explanation
Correct Answer
D:Inspect the actual packet path, firewall policy and service identity; review a least-privilege change before deployment. A concrete inspection identifies the destination, port, path and effective firewall controls. Change control then evaluates a restricted rule.
Incorrect Answers
A:Apply the generated rule directly because assistant suggestions always override network security requirements. Generated suggestions are not proof of a correct or authorized configuration. A world-open rule risks unrelated services and must be reviewed.
B:Give every workload account organization Owner so the assistant no longer needs to inspect firewall policies. Broad organization access cannot establish which network path is failing; it would introduce a severe IAM exposure instead.
C:Disable Firewall Rules Logging, which validates that the generated ingress change is safe for every target. Logging controls observation and audit evidence; disabling it does not validate the necessity or scope of a firewall rule.
Question 114
A project has an unexpected interservice network path and a teammate wants to use Gemini Cloud Assist to understand the existing layout. Which input should accompany the request?
- Provide only an invented architecture diagram without verifying live project data.
- Grant the assistant unrestricted Owner access to every project to improve its analysis.
- Provide authorized project and relevant network context, then corroborate recommendations using current resource settings.
- Ask the assistant to infer the network topology from the project’s billing budget.
Explanation
Correct Answer
C:Resource identifiers, VPC/subnet details, routes and applicable firewall policies help frame the actual problem. Compare the assistant’s answer with current configuration and observed packet paths before using its guidance for changes.
Incorrect Answers
A:An out-of-date or hypothetical diagram is insufficient evidence of the paths currently configured.
B:Broad IAM escalation is unnecessary for a focused diagnostic and violates least privilege.
D:A billing budget has cost data, not the authoritative map of network paths.
Question 115
Gemini Cloud Assist suggests that project atlas has 800 regional CPUs available, but its answer provides no metric, region, or observation time. Before increasing a production batch fleet in europe-west4, what evidence must the operator obtain?
- Increase the cloud billing budget because budget headroom raises regional Compute Engine CPU quota automatically.
- Treat the AI response as the authoritative quota value because it was generated from a Google-hosted tool.
- Read the actual quota metric and current utilization for project atlas and europe-west4 in Quotas or the supported quota API; also check capacity.
- Use the CPU quota from a development project in us-central1 because quota applies uniformly to all projects and regions.
Explanation
Correct Answer
C:Read the actual quota metric and current utilization for project atlas and europe-west4 in Quotas or the supported quota API; also check capacity. Google quotas are scoped by resource metric and often by region or other dimensions. The project-and-region quota view directly tests the number the assistant omitted; quota also does not guarantee zonal machine availability.
Incorrect Answers
A:Increase the cloud billing budget because budget headroom raises regional Compute Engine CPU quota automatically. Billing limits spending notifications and payment responsibility; it does not increase a project-specific regional CPU quota.
B:Treat the AI response as the authoritative quota value because it was generated from a Google-hosted tool. An assistant-generated number without a scope or timestamp cannot prove an actual quota allocation. Production capacity decisions require authoritative quota data.
D:Use the CPU quota from a development project in us-central1 because quota applies uniformly to all projects and regions. A quota for another project or region cannot establish headroom for atlas in europe-west4, even if its machine family is the same.
Question 116
An AI chat answer says VM api-4 still has a public IPv4 address, but the diagram it cites is two weeks old. The on-call engineer needs its present network exposure before a firewall change. Which check is authoritative?
- Use the VM diagram from the design document, because an approved diagram always tracks the current NIC configuration.
- Inspect api-4 in Compute Engine or its current instance API record and check networkInterfaces.accessConfigs for an external IPv4 assignment.
- Examine only the project's billing export; a publicly addressable VM is always identifiable by a separate internet egress cost line.
- Query a DNS resolver for the application domain and assume the returned address proves that api-4 itself has an external interface.
Explanation
Correct Answer
B:Inspect api-4 in Compute Engine or its current instance API record and check networkInterfaces.accessConfigs for an external IPv4 assignment. Compute Engine stores the currently attached network interface and external access configuration. That live resource record answers whether this VM currently has an external IPv4 address.
Incorrect Answers
A:Use the VM diagram from the design document, because an approved diagram always tracks the current NIC configuration. Diagrams can lag behind instance updates; the question asks for present VM interface configuration rather than intended architecture.
C:Examine only the project's billing export; a publicly addressable VM is always identifiable by a separate internet egress cost line. Billing usage is aggregated and delayed, and an egress charge does not identify whether the VM has a public IP.
D:Query a DNS resolver for the application domain and assume the returned address proves that api-4 itself has an external interface. A public DNS record can point to a load balancer or another endpoint. It does not prove the specific backend VM has an external access config.
Question 117
An AI-assisted remediation proposes expanding a production IAM role to solve one read failure. Which safer next step aligns with least privilege?
- Inspect the denied operation and grant only the needed permission through an approved minimal role or scoped binding.
- Accept the assistant’s recommendation and grant project Owner without reviewing the request.
- Disable IAM audit logging to prevent additional access-denied alerts.
- Add the caller to an organization-wide administrator group for the duration of the task.
Explanation
Correct Answer
A:First identify the specific permission and target resource from the failure. A minimal supported IAM role or reviewed custom role at the correct scope resolves the need without treating an assistant suggestion as authorization to escalate broadly.
Incorrect Answers
B:Owner adds unrelated privileges and may not address the actual authorization boundary safely.
C:Suppressing audit logs hides evidence and has no effect on the denied permission.
D:Organization-wide membership exceeds the narrowly scoped read operation and increases security exposure.
Question 118
Gemini Cloud Assist summarizes resource availability across two regions, but one planned service is not yet offered in either target region. How should a deployment decision be made?
- Choose either planned region based solely on a natural-language assistant summary.
- Create the service in an unsupported region and assume Google Cloud will transparently move it.
- Check the current product-specific region/feature availability and adapt the regional architecture accordingly.
- Move the billing account to the country where the product is offered.
Explanation
Correct Answer
C:Service features can launch by region and change over time. Confirm both the target service and the exact required feature are supported in the planned locations, then select supported regions or redesign the resilience plan.
Incorrect Answers
A:An assistant summary is not a reliable substitute for current product feature availability by region.
B:A deployment that specifies an unsupported region or capability may be rejected; automatic relocation cannot be assumed.
D:Billing account administration does not make a service feature available in a new geographic region.
Question 119
An assistant suggests deleting an idle resource based on low utilization, but the VM is the only active disaster-recovery standby. Which check should block deletion?
- Delete the idle standby immediately because zero CPU usage proves it is unused.
- Reboot it and then delete it without speaking with the recovery owner.
- Confirm its recovery role, required readiness, dependencies and approved change path before any removal.
- Create an alerts-only billing budget and use the budget amount to determine disaster recovery coverage.
Explanation
Correct Answer
C:The team must assess recovery objectives and dependencies, not just average utilization. Validate the standby’s role with the service owner and preserve an approved replacement or rollback before deleting any disaster-recovery component.
Incorrect Answers
A:Standby systems are often intentionally idle and may be essential during an outage.
B:A restart does not establish whether the resource is safe to retire or fulfill change controls.
D:A spending budget does not determine whether the recovery design remains operational.
Question 120
A team wants to document the source of a Gemini Cloud Assist troubleshooting recommendation for an incident ticket. What must the engineer preserve?
- Copy only the assistant response and treat the generated text as its own independent citation.
- Preserve the project name but discard the errors and settings that led to the recommendation.
- Record the configuration, logs, evidence timestamps, actions taken and checks that confirmed the conclusion.
- Document only the virtual-machine instance size since incidents are always capacity related.
Explanation
Correct Answer
C:An incident record should connect the hypothesis to observed errors, relevant resource state and verified remediation outcomes. Preserve enough timestamped context for another engineer to understand and reproduce the decision.
Incorrect Answers
A:An AI-generated recommendation can be mistaken and needs corroborating operational evidence.
B:A project name alone cannot reproduce the diagnosis or distinguish later configuration changes.
D:Incidents may involve IAM, networking or application behavior, not just instance capacity.