100% Latest & Updated Google Professional Cloud Architect Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!
Question 390
A platform team needs a quality-control process for new managed infrastructure templates before developers may self-service them. Which implementation choice most directly delivers automated template validation?
- Skip disposable test deployments and promote straight to production
- Give template tests permanent Owner privileges in target projects
- Run policy and integration tests on disposable environments before catalog promotion
- Ignore rollback behavior when a module update fails
Explanation
Correct Answer
C:A valid configuration file can still deploy an insecure or unusable resource
Incorrect Answers
A:Not integration validation. A valid configuration file can still deploy an insecure or unusable resource.
B:Unnecessary and dangerous. A valid configuration file can still deploy an insecure or unusable resource.
D:Not deployability control. A valid configuration file can still deploy an insecure or unusable resource.
Question 391
A platform team needs a quality-control process for new managed infrastructure templates before developers may self-service them. Which validation would best establish proof of quality controls rejecting defects?
- Test only valid templates with no deliberate risky grant
- Inspect documentation but never exercise policy engine
- Introduce a deliberate excessive IAM grant and verify template promotion is blocked
- Check deployment speed while leaving unsafe IAM undetected
Explanation
Correct Answer
C:Positive tests alone do not prove unsafe infrastructure is prevented
Incorrect Answers
A:Cannot prove negative gate. Positive tests alone do not prove unsafe infrastructure is prevented.
B:Not enforcement evidence. Positive tests alone do not prove unsafe infrastructure is prevented.
D:Wrong outcome. Positive tests alone do not prove unsafe infrastructure is prevented.
Question 392
A platform team needs a quality-control process for new managed infrastructure templates before developers may self-service them. Which operational action best maintains periodic control effectiveness review after launch?
- Measure escaped template defects and update automated checks from incidents
- Treat any escaped defect as user error and leave tests unchanged
- Monitor only template download counts
- Delete failed promotion logs before root cause analysis
Explanation
Correct Answer
A:Quality controls must adapt as products and usage change
Incorrect Answers
B:Quality process does not learn. Quality controls must adapt as products and usage change.
C:Not defects. Quality controls must adapt as products and usage change.
D:Loses correction evidence. Quality controls must adapt as products and usage change.
Question 393
A customer booking service has redundant zones but its shared database connections become exhausted during sudden traffic spikes. Which architecture most directly meets the need for reliability at dependency boundaries?
- Add more API replicas without bounding shared database connections
- Use one larger database VM without backpressure
- Design backpressure and bounded connection pooling around the shared database
- Rely solely on Cloud CDN for authenticated write requests
Explanation
Correct Answer
C:Redundant application replicas cannot help if they all overload the same dependency
Incorrect Answers
A:May worsen dependency overload. Redundant application replicas cannot help if they all overload the same dependency.
B:Can still fail during spikes. Redundant application replicas cannot help if they all overload the same dependency.
D:Does not protect database writes. Redundant application replicas cannot help if they all overload the same dependency.
Question 394
A customer booking service has redundant zones but its shared database connections become exhausted during sudden traffic spikes. Which implementation choice most directly delivers overload protection implementation?
- Allow each request to open an unbounded new DB connection
- Retry failed queries instantly without budget or jitter
- Set connection pool limits, timeouts and safe retry budgets for API workers
- Configure timeouts longer than customer SLO with no circuit breaking
Explanation
Correct Answer
C:Unlimited retries can multiply database overload during partial failures
Incorrect Answers
A:Saturation worsens. Unlimited retries can multiply database overload during partial failures.
B:Can create a retry storm. Unlimited retries can multiply database overload during partial failures.
D:Work piles up under failure. Unlimited retries can multiply database overload during partial failures.
Question 395
A customer booking service has redundant zones but its shared database connections become exhausted during sudden traffic spikes. Which validation would best establish proof of graceful overload?
- Load test past the saturation point and verify controlled errors instead of cascading failure
- Test only below saturation threshold
- Inspect CPU at normal traffic
- Measure success requests without rejected/error rate during spikes
Explanation
Correct Answer
A:High throughput at normal load does not prove stable behavior during overload
Incorrect Answers
B:Does not reveal overload behavior. High throughput at normal load does not prove stable behavior during overload.
C:Not graceful failure evidence. High throughput at normal load does not prove stable behavior during overload.
D:Can hide cascading failure. High throughput at normal load does not prove stable behavior during overload.
Question 396
A customer booking service has redundant zones but its shared database connections become exhausted during sudden traffic spikes. Which operational action best maintains continuous dependency capacity review after launch?
- Monitor only frontend replica count
- Increase max autoscale instances whenever queries fail
- Ignore queue latency until requests time out
- Monitor active connections and queue latency and tune limits as traffic patterns change
Explanation
Correct Answer
D:Dependency saturation can recur even when autoscaling adds application instances
Incorrect Answers
A:Not shared DB saturation. Dependency saturation can recur even when autoscaling adds application instances.
B:May amplify connections. Dependency saturation can recur even when autoscaling adds application instances.
C:Intervene too late. Dependency saturation can recur even when autoscaling adds application instances.
Question 397
A health-record API must survive an availability-zone fault and prove that its external users can still complete critical requests. Which architecture most directly meets the need for multi-zone business continuity?
- Keep all serving replicas in one availability zone
- Replicate data in another region but leave all API dependencies in one zone
- Deploy independent serving zones and ensure the necessary data tier also tolerates zonal loss
- Rely on a manual restart runbook without spare capacity
Explanation
Correct Answer
C:Replicated application nodes are inadequate if the data tier is a single-zone dependency
Incorrect Answers
A:Single-zone failure still removes service. Replicated application nodes are inadequate if the data tier is a single-zone dependency.
B:Incomplete zonal tolerance. Replicated application nodes are inadequate if the data tier is a single-zone dependency.
D:May fail fast restoration. Replicated application nodes are inadequate if the data tier is a single-zone dependency.
Question 398
A health-record API must survive an availability-zone fault and prove that its external users can still complete critical requests. Which implementation choice most directly delivers ready failover capacity?
- Add second-zone instances without health-based routing
- Assign all capacity to primary zone with no reserve
- Treat storage backups as a replacement for healthy live DB failover
- Set load-balancer health checks and reserve capacity in surviving zones
Explanation
Correct Answer
D:Traffic can fail despite redundant zones when backup capacity is not available
Incorrect Answers
A:Traffic can continue going to failed nodes. Traffic can fail despite redundant zones when backup capacity is not available.
B:Secondary cannot meet load. Traffic can fail despite redundant zones when backup capacity is not available.
C:Not online service continuity. Traffic can fail despite redundant zones when backup capacity is not available.
Question 399
A health-record API must survive an availability-zone fault and prove that its external users can still complete critical requests. Which validation would best establish proof of end-to-end regional-zone resilience?
- Ping the surviving node after zone loss
- Check only cloud resource health status
- Measure VM boot times and ignore authenticated API operations
- Stop one zone during a test and complete authenticated patient record reads and writes
Explanation
Correct Answer
D:A VM ping does not establish the clinical workflow is still functioning
Incorrect Answers
A:Does not prove patient read/write success. A VM ping does not establish the clinical workflow is still functioning.
B:Not clinical workflow. A VM ping does not establish the clinical workflow is still functioning.
C:Not business recovery. A VM ping does not establish the clinical workflow is still functioning.
Question 400
A health-record API must survive an availability-zone fault and prove that its external users can still complete critical requests. Which operational action best maintains operational reliability exercises after launch?
- Stop running recovery drills after first success
- Monitor only monthly infrastructure spend instead of transaction SLOs
- Schedule controlled failover drills and track time to restore clinical transaction SLOs
- Ignore failed health checks as long as some VMs are running
Explanation
Correct Answer
C:Failover configuration changes can erode reliability without repeated practical validation
Incorrect Answers
A:Future changes may break failover. Failover configuration changes can erode reliability without repeated practical validation.
B:Cannot detect resilience regressions. Failover configuration changes can erode reliability without repeated practical validation.
D:May conceal unavailable client routes. Failover configuration changes can erode reliability without repeated practical validation.
Google Cloud Certified - Professional Cloud Architect (Professional Cloud Architect) Practice Tests