Google Professional Cloud Architect Practice Test 60
Question 590
At National grocery cooperative, the Google Cloud design for Gemini Cloud Assist runbook recommendations is affected because an engineer proposes applying suggested infrastructure edits directly to production. A review found that review finds recommendations do not reflect a special pharmacy-workload SLA. The proposed response must account for the requirement that critical pharmacy distribution cannot be disrupted by unreviewed changes. The cause has been established and a lasting architecture is being selected. Which design best meets the constraint?
- Validate AI-assisted suggestions against live architecture, policy and business SLOs
- Execute every model-generated shell command as project Owner without review
- Track AI-assisted changes, rejected suggestions and postrelease incidents
- Define human-reviewed change proposals with policy and unit tests
Explanation
Correct Answer
D:
Define human-reviewed change proposals with policy and unit tests targets the underlying issue (an engineer proposes applying suggested infrastructure edits directly to production) rather than merely reporting symptoms. It also preserves the required outcome (critical pharmacy distribution cannot be disrupted by unreviewed changes), which a target-state design must satisfy.
Incorrect Answers
A:
The alternative “Validate AI-assisted suggestions against live architecture, policy and business SLOs” investigates the cause without specifying the durable architecture to resolve it. In the reported observation, review finds recommendations do not reflect a special pharmacy-workload SLA. The option “Define human-reviewed change proposals with policy and unit tests” instead addresses the requested architecture selection.
B:
An AI suggestion is not an authorized or validated operational change. For this question, it is not a defensible choice for architecture selection and it does not satisfy the requirement that critical pharmacy distribution cannot be disrupted by unreviewed changes.
C:
The alternative “Track AI-assisted changes, rejected suggestions and postrelease incidents” measures results without defining the preventive design now required. In the reported observation, review finds recommendations do not reflect a special pharmacy-workload SLA. The option “Define human-reviewed change proposals with policy and unit tests” instead addresses the requested architecture selection.
Question 591
At University public data project, the sensitive research boundary workload has a documented risk: the analytics team shares projects with public application workloads. The review found that review finds protected training datasets readable from an unrelated compute project. The architecture must also ensure that university policy forbids exfiltration of patient-linked samples. The team must deliver the approved control safely. Which deployment sequence is most appropriate?
- Stage perimeter dry-run and egress policies before enforcing access boundaries
- Use only object naming conventions to stop exfiltration
- Review perimeter violations, allowed egress exemptions and new projects
- Use VPC Service Controls perimeters around supported sensitive services along with IAM
Explanation
Correct Answer
A:
Stage perimeter dry-run and egress policies before enforcing access boundaries is a controlled means to introduce the approved architecture while meeting the requirement that university policy forbids exfiltration of patient-linked samples. Staged delivery, attributable change evidence and rollback guard against an otherwise avoidable service disruption.
Incorrect Answers
B:
Naming cannot enforce data-access boundaries. For this question, it is not a defensible choice for controlled implementation and it does not satisfy the requirement that university policy forbids exfiltration of patient-linked samples.
C:
The alternative “Review perimeter violations, allowed egress exemptions and new projects” addresses later monitoring rather than the actual controlled deployment. In the reported observation, review finds protected training datasets readable from an unrelated compute project. The option “Stage perimeter dry-run and egress policies before enforcing access boundaries” instead addresses the requested controlled implementation.
D:
The alternative “Use VPC Service Controls perimeters around supported sensitive services along with IAM” describes the chosen design, but not how it is introduced safely. In the reported observation, review finds protected training datasets readable from an unrelated compute project. The option “Stage perimeter dry-run and egress policies before enforcing access boundaries” instead addresses the requested controlled implementation.
Question 592
Case study: Online ticket lottery: reservation API
Online ticket lottery relies on reservation API to deliver its operational commitments. The current architectural problem is that the planned single service is sized for ordinary traffic but purchases arrive in minute-long surges. The team has recorded this evidence: load tests show 40x burst with a limited pool of unique seats. Any proposed response must preserve the requirement that a seat must not be reserved twice under concurrent requests.
At Online ticket lottery, the Google Cloud design for reservation API is affected because the planned single service is sized for ordinary traffic but purchases arrive in minute-long surges. A review found that load tests show 40x burst with a limited pool of unique seats. The proposed response must account for the requirement that a seat must not be reserved twice under concurrent requests. A deployment ticket has been closed. What direct evidence would justify solution signoff?
- Cache seat ownership only in eventually consistent edge storage
- Canary the new path with bounded traffic and rollback under load
- Simulate 40x competing requests and prove one paid allocation per seat
- Measure burst concurrency, lock contention and reservation expiry windows
Explanation
Correct Answer
C:
Simulate 40x competing requests and prove one paid allocation per seat is a direct acceptance test of the path implicated by load tests show 40x burst with a limited pool of unique seats. It checks that the original risk (the planned single service is sized for ordinary traffic but purchases arrive in minute-long surges) has actually been controlled and tests that permitted operations still meet the requirement that a seat must not be reserved twice under concurrent requests.
Incorrect Answers
A:
Stale edge ownership cannot prevent double booking. For this question, it is not a defensible choice for acceptance verification and it does not satisfy the requirement that a seat must not be reserved twice under concurrent requests.
B:
The alternative “Canary the new path with bounded traffic and rollback under load” documents change execution but not its operational acceptance result. In the reported observation, load tests show 40x burst with a limited pool of unique seats. The option “Simulate 40x competing requests and prove one paid allocation per seat” instead addresses the requested acceptance verification.
D:
The alternative “Measure burst concurrency, lock contention and reservation expiry windows” collects prechange evidence without confirming that the revised service meets its outcome. In the reported observation, load tests show 40x burst with a limited pool of unique seats. The option “Simulate 40x competing requests and prove one paid allocation per seat” instead addresses the requested acceptance verification.
Question 593
The private AI endpoint access environment at Global pharmacy vendor faces this issue: public model endpoints are reachable by unauthorized IPs while containing protected pricing prompts. Supporting evidence shows that network scan finds broad ingress on a model-serving load balancer. The architect must preserve the requirement that only corporate approved callers may submit sensitive requests. Go-live and initial tests passed. Which ongoing measure is most useful to catch control drift or regression?
- Monitor ingress drift and attempted model endpoint abuse
- Hide endpoint names in source code but leave public access open
- Use private model-serving access and workload-bound IAM with constrained ingress
- Try unauthorized public access and confirm private authenticated clients still work
Explanation
Correct Answer
A:
Monitor ingress drift and attempted model endpoint abuse reveals recurrence, drift or changing dependencies after launch. It provides a repeatable signal for the risk (public model endpoints are reachable by unauthorized IPs while containing protected pricing prompts) rather than treating the first successful test as permanent assurance.
Incorrect Answers
B:
Obscurity is not an endpoint authorization control. For this question, it is not a defensible choice for ongoing operational assurance and it does not satisfy the requirement that only corporate approved callers may submit sensitive requests.
C:
The alternative “Use private model-serving access and workload-bound IAM with constrained ingress” selects a design, but does not provide the ongoing signal required to detect drift. In the reported observation, network scan finds broad ingress on a model-serving load balancer. The option “Monitor ingress drift and attempted model endpoint abuse” instead addresses the requested ongoing operational assurance.
D:
The alternative “Try unauthorized public access and confirm private authenticated clients still work” is a one-time acceptance result rather than ongoing operational assurance. In the reported observation, network scan finds broad ingress on a model-serving load balancer. The option “Monitor ingress drift and attempted model endpoint abuse” instead addresses the requested ongoing operational assurance.
Question 594
Consider Solar equipment company and its predictive maintenance model workload. The issue is that trained model artifacts are deployed manually and inference behavior drifts across regions; the current observation is that model logs show differing preprocessing versions between plant gateways. The team must ensure that anomaly decisions must have lineage and allow rollback. The change committee needs to understand the current system rather than install a fix yet. What should it do?
- Edit model weights directly inside each running inference container
- Deploy canary model versions with signed artifacts and rollback gate
- Inventory feature pipeline versions, model registry lineage and inference paths
- Use Vertex AI model registry and versioned serving with consistent preprocessing
Explanation
Correct Answer
C:
Inventory feature pipeline versions, model registry lineage and inference paths is the appropriate diagnostic because the observation—model logs show differing preprocessing versions between plant gateways—does not by itself establish every dependency, authorization rule or failure boundary. The findings explain the reported risk (trained model artifacts are deployed manually and inference behavior drifts across regions) before the design is approved.
Incorrect Answers
A:
Unversioned manual edits destroy reproducibility and rollback. For this question, it is not a defensible choice for root-cause evidence and it does not satisfy the requirement that anomaly decisions must have lineage and allow rollback.
B:
The alternative “Deploy canary model versions with signed artifacts and rollback gate” is rollout work that may modify the wrong dependency before root-cause evidence is collected. In the reported observation, model logs show differing preprocessing versions between plant gateways. The option “Inventory feature pipeline versions, model registry lineage and inference paths” instead addresses the requested root-cause evidence.
D:
The alternative “Use Vertex AI model registry and versioned serving with consistent preprocessing” is target architecture selection; it does not establish why the observed path failed. In the reported observation, model logs show differing preprocessing versions between plant gateways. The option “Inventory feature pipeline versions, model registry lineage and inference paths” instead addresses the requested root-cause evidence.
Question 595
At International shipping brokerage, the Google Cloud design for shipment tracking platform is affected because executives want real-time tracking without defining the latency and availability targets. A review found that port partners deliver events with variable delays and occasional duplicate identifiers. The proposed response must account for the requirement that customs declarations require nonduplicated status transitions. The solution is operating. Which evidence should be watched for longer-term reliability and compliance?
- Inject late and duplicate events and verify the customs status remains consistent
- Assume ordered, exactly-once delivery from every external partner without validation
- Architect idempotent event ingestion and a separately materialized customer tracking view
- Measure lag per partner, invalid transitions and delivery-cost trends
Explanation
Correct Answer
D:
Measure lag per partner, invalid transitions and delivery-cost trends reveals recurrence, drift or changing dependencies after launch. It provides a repeatable signal for the risk (executives want real-time tracking without defining the latency and availability targets) rather than treating the first successful test as permanent assurance.
Incorrect Answers
A:
The alternative “Inject late and duplicate events and verify the customs status remains consistent” is a one-time acceptance result rather than ongoing operational assurance. In the reported observation, port partners deliver events with variable delays and occasional duplicate identifiers. The option “Measure lag per partner, invalid transitions and delivery-cost trends” instead addresses the requested ongoing operational assurance.
B:
External partner delivery can be duplicated or out of order. For this question, it is not a defensible choice for ongoing operational assurance and it does not satisfy the requirement that customs declarations require nonduplicated status transitions.
C:
The alternative “Architect idempotent event ingestion and a separately materialized customer tracking view” selects a design, but does not provide the ongoing signal required to detect drift. In the reported observation, port partners deliver events with variable delays and occasional duplicate identifiers. The option “Measure lag per partner, invalid transitions and delivery-cost trends” instead addresses the requested ongoing operational assurance.
Question 596
Consider Legal case management SaaS and its multi-tenant object security workload. The issue is that valid user tokens can access case files belonging to other law firms; the current observation is that penetration test exchanges a case ID and retrieves another tenant's document. The team must ensure that each firm requires strict object-level isolation. The target architecture is approved; execution without disruption is the remaining decision. Which rollout action is best?
- Introduce safe policy fixes with a dual-tenant regression suite
- Apply tenant-bound object authorization enforced on every API and data query
- Monitor denied cross-tenant access and new API routes lacking ownership checks
- Rely only on login authentication without checking case object ownership
Explanation
Correct Answer
A:
Introduce safe policy fixes with a dual-tenant regression suite is a controlled means to introduce the approved architecture while meeting the requirement that each firm requires strict object-level isolation. Staged delivery, attributable change evidence and rollback guard against an otherwise avoidable service disruption.
Incorrect Answers
B:
The alternative “Apply tenant-bound object authorization enforced on every API and data query” describes the chosen design, but not how it is introduced safely. In the reported observation, penetration test exchanges a case ID and retrieves another tenant's document. The option “Introduce safe policy fixes with a dual-tenant regression suite” instead addresses the requested controlled implementation.
C:
The alternative “Monitor denied cross-tenant access and new API routes lacking ownership checks” addresses later monitoring rather than the actual controlled deployment. In the reported observation, penetration test exchanges a case ID and retrieves another tenant's document. The option “Introduce safe policy fixes with a dual-tenant regression suite” instead addresses the requested controlled implementation.
D:
Authentication cannot replace per-object authorization. For this question, it is not a defensible choice for controlled implementation and it does not satisfy the requirement that each firm requires strict object-level isolation.
Question 597
Biopharma laboratory plans to update internal research dashboards. Investigators found that workload trace shows long scans that lock operational assay tables, consistent with the reported problem that analysts need petabyte SQL aggregation but a transactional app database is overloaded. The response must satisfy the requirement that regulated sample updates cannot be blocked by ad hoc reporting. The approved change has been applied. Which validation provides the clearest evidence the original problem is fixed?
- Point every BI user at the operational database with unrestricted full-table scans
- Compare query aggregates and measure impact on the operational database
- Schedule partitioned incremental extraction and dataset-level controls
- Differentiate OLTP writes, query concurrency and analytics freshness
Explanation
Correct Answer
B:
Compare query aggregates and measure impact on the operational database is a direct acceptance test of the path implicated by workload trace shows long scans that lock operational assay tables. It checks that the original risk (analysts need petabyte SQL aggregation but a transactional app database is overloaded) has actually been controlled and tests that permitted operations still meet the requirement that regulated sample updates cannot be blocked by ad hoc reporting.
Incorrect Answers
A:
Unbounded analytics queries undermine transactional service. For this question, it is not a defensible choice for acceptance verification and it does not satisfy the requirement that regulated sample updates cannot be blocked by ad hoc reporting.
C:
The alternative “Schedule partitioned incremental extraction and dataset-level controls” documents change execution but not its operational acceptance result. In the reported observation, workload trace shows long scans that lock operational assay tables. The option “Compare query aggregates and measure impact on the operational database” instead addresses the requested acceptance verification.
D:
The alternative “Differentiate OLTP writes, query concurrency and analytics freshness” collects prechange evidence without confirming that the revised service meets its outcome. In the reported observation, workload trace shows long scans that lock operational assay tables. The option “Compare query aggregates and measure impact on the operational database” instead addresses the requested acceptance verification.
Question 598
At Automotive warranty platform, the Google Cloud design for parts inventory read API is affected because a global rollout uses one master inventory database for every dealership lookup. A review found that traces show cross-ocean reads dominate response time. The proposed response must account for the requirement that inventory commitments must remain serializable where parts are scarce. Which controlled implementation should be used to preserve operational continuity during the change?
- Use independent writable inventory databases without conflict handling
- Create region-proximate read views while keeping authoritative reservation writes consistent
- Track replica lag, order conflicts and cross-region read costs
- Pilot regional read distribution with explicit staleness indicators
Explanation
Correct Answer
D:
Pilot regional read distribution with explicit staleness indicators is a controlled means to introduce the approved architecture while meeting the requirement that inventory commitments must remain serializable where parts are scarce. Staged delivery, attributable change evidence and rollback guard against an otherwise avoidable service disruption.
Incorrect Answers
A:
Uncoordinated writable replicas can oversell scarce parts. For this question, it is not a defensible choice for controlled implementation and it does not satisfy the requirement that inventory commitments must remain serializable where parts are scarce.
B:
The alternative “Create region-proximate read views while keeping authoritative reservation writes consistent” describes the chosen design, but not how it is introduced safely. In the reported observation, traces show cross-ocean reads dominate response time. The option “Pilot regional read distribution with explicit staleness indicators” instead addresses the requested controlled implementation.
C:
The alternative “Track replica lag, order conflicts and cross-region read costs” addresses later monitoring rather than the actual controlled deployment. In the reported observation, traces show cross-ocean reads dominate response time. The option “Pilot regional read distribution with explicit staleness indicators” instead addresses the requested controlled implementation.
Question 599
The Apigee API consumer management environment at Regional airline booking platform faces this issue: travel agencies call the same reservation endpoint with no consumer-specific quota. Supporting evidence shows that gateway logs show one agency's retry loop overwhelms others. The architect must preserve the requirement that partners need predictable access without shutting down public bookings. Before approving a solution, which analysis most directly clarifies the underlying dependency or control gap?
- Identify API consumers, quota profiles and retry error causes
- Apply one secret shared by every travel agency with no rate controls
- Introduce staged quota enforcement in preview and notify partners
- Design Apigee API products with per-consumer credentials, quotas and analytics
Explanation
Correct Answer
A:
Identify API consumers, quota profiles and retry error causes is the appropriate diagnostic because the observation—gateway logs show one agency's retry loop overwhelms others—does not by itself establish every dependency, authorization rule or failure boundary. The findings explain the reported risk (travel agencies call the same reservation endpoint with no consumer-specific quota) before the design is approved.
Incorrect Answers
B:
Shared identity and unrestricted rate make abusive tenants unaccountable. For this question, it is not a defensible choice for root-cause evidence and it does not satisfy the requirement that partners need predictable access without shutting down public bookings.
C:
The alternative “Introduce staged quota enforcement in preview and notify partners” is rollout work that may modify the wrong dependency before root-cause evidence is collected. In the reported observation, gateway logs show one agency's retry loop overwhelms others. The option “Identify API consumers, quota profiles and retry error causes” instead addresses the requested root-cause evidence.
D:
The alternative “Design Apigee API products with per-consumer credentials, quotas and analytics” is target architecture selection; it does not establish why the observed path failed. In the reported observation, gateway logs show one agency's retry loop overwhelms others. The option “Identify API consumers, quota profiles and retry error causes” instead addresses the requested root-cause evidence.
Question 600
Consider City parking payment app and its payment token audit workload. The issue is that service records full PAN in debug logs while application database is tokenized; the current observation is that log sample includes unmasked card numbers in error traces. The team must ensure that payment operations require minimized log retention and incident-ready evidence. Which continuous monitoring approach best reveals recurrence or changing dependencies?
- Monitor new PAN detections and changes in logging pipeline
- Use TLS on log transport and keep full card numbers indefinitely
- Remove PAN from telemetry and apply scoped tokenization and log redaction
- Generate test errors and verify no PAN appears in log destinations
Explanation
Correct Answer
A:
Monitor new PAN detections and changes in logging pipeline reveals recurrence, drift or changing dependencies after launch. It provides a repeatable signal for the risk (service records full PAN in debug logs while application database is tokenized) rather than treating the first successful test as permanent assurance.
Incorrect Answers
B:
Transport encryption alone does not minimize sensitive logged content. For this question, it is not a defensible choice for ongoing operational assurance and it does not satisfy the requirement that payment operations require minimized log retention and incident-ready evidence.
C:
The alternative “Remove PAN from telemetry and apply scoped tokenization and log redaction” selects a design, but does not provide the ongoing signal required to detect drift. In the reported observation, log sample includes unmasked card numbers in error traces. The option “Monitor new PAN detections and changes in logging pipeline” instead addresses the requested ongoing operational assurance.
D:
The alternative “Generate test errors and verify no PAN appears in log destinations” is a one-time acceptance result rather than ongoing operational assurance. In the reported observation, log sample includes unmasked card numbers in error traces. The option “Monitor new PAN detections and changes in logging pipeline” instead addresses the requested ongoing operational assurance.