100% Latest & Updated Google Professional Cloud Architect Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!
Question 170
An enterprise searches policy manuals and needs governed answers with traceable citations rather than raw document excerpts. Which implementation choice most directly delivers verifiable policy-source citations?
- Paste a manually chosen source title next to the final answer
- Use embedding similarity scores as the citation shown to auditors
- Store generated answers without passage IDs or document versions
- Configure retrieval to retain passage identifiers, source versions, and citation metadata for each generated answer
Explanation
Correct Answer
D:A traceable answer points to the precise approved policy source rather than an unverifiable model summary
Incorrect Answers
A:Does not bind a claim to the actual retrieved passage or policy version. A traceable answer points to the precise approved policy source rather than an unverifiable model summary.
B:Similarity scores are not verifiable policy sources. A traceable answer points to the precise approved policy source rather than an unverifiable model summary.
C:Reviewers cannot trace a claim to the approved underlying document. A traceable answer points to the precise approved policy source rather than an unverifiable model summary.
Question 171
An enterprise searches policy manuals and needs governed answers with traceable citations rather than raw document excerpts. Which validation would best establish proof of cited answer reliability?
- Run labeled questions and test citation accuracy and private-document refusal
- Ask users whether answers feel helpful without checking citations
- Test only public documents while ignoring confidential policy files
- Check ingestion speed alone
Explanation
Correct Answer
A:A fluent answer may be unsupported or expose content from another employee group
Incorrect Answers
B:Does not prove grounding. A fluent answer may be unsupported or expose content from another employee group.
C:Misses access control issues. A fluent answer may be unsupported or expose content from another employee group.
D:Does not verify relevance or safe refusal. A fluent answer may be unsupported or expose content from another employee group.
Question 172
An enterprise searches policy manuals and needs governed answers with traceable citations rather than raw document excerpts. Which operational action best maintains continuous answer-quality stewardship after launch?
- Ignore newly revoked document permissions until annual reindex
- Monitor retrieval relevance, index freshness, and source permission changes
- Monitor only model API uptime
- Promote unreviewed index changes without sampling answer attribution
Explanation
Correct Answer
B:Stale indexes and outdated ACLs can erode trust despite unchanged application code
Incorrect Answers
A:Can leak old access. Stale indexes and outdated ACLs can erode trust despite unchanged application code.
C:Misses retrieval relevance regressions. Stale indexes and outdated ACLs can erode trust despite unchanged application code.
D:Can introduce unsupported outputs. Stale indexes and outdated ACLs can erode trust despite unchanged application code.
Question 173
A platform team has been granting project Editor permissions to application developers who only need to deploy one service. Which architecture most directly meets the need for least-privilege project access?
- Keep project Editor for all developers because one service needs deployment rights
- Use a single shared administrator identity for all development teams
- Define service-specific predefined or custom IAM roles at the smallest required scope
- Grant organization Owner as a shortcut to project deployment
Explanation
Correct Answer
C:Broad project Editor grants unrelated mutation rights and enlarge the blast radius
Incorrect Answers
A:Grants unnecessary unrelated resource privileges. Broad project Editor grants unrelated mutation rights and enlarge the blast radius.
B:Removes individual accountability. Broad project Editor grants unrelated mutation rights and enlarge the blast radius.
D:Widens access far beyond the service. Broad project Editor grants unrelated mutation rights and enlarge the blast radius.
Question 174
A platform team has been granting project Editor permissions to application developers who only need to deploy one service. Which implementation choice most directly delivers controlled developer role assignment?
- Assign deployment grants to individual accounts without lifecycle group management
- Let developers self-approve new IAM role bindings
- Bind deployment permissions to named groups and short-lived identities, not shared accounts
- Use an unrestricted access token copied between team members
Explanation
Correct Answer
C:Group-based accountable grants can be reviewed and removed consistently
Incorrect Answers
A:Makes offboarding inconsistent. Group-based accountable grants can be reviewed and removed consistently.
B:Permits unchecked privilege growth. Group-based accountable grants can be reviewed and removed consistently.
D:Cannot maintain attributable permissions. Group-based accountable grants can be reviewed and removed consistently.
Question 175
A platform team has been granting project Editor permissions to application developers who only need to deploy one service. Which validation would best establish proof of excessive privilege removal?
- Check developers can still deploy the service and stop there
- Use policy analysis and denied-action tests for unrelated sensitive operations
- Inspect the number of IAM roles but ignore their effective permissions
- Run a cloud resource inventory but never test as a developer
Explanation
Correct Answer
B:Login success alone cannot show that developers lack unrelated administrative powers
Incorrect Answers
A:Does not test denied unrelated operations. Login success alone cannot show that developers lack unrelated administrative powers.
C:Role count is not least privilege. Login success alone cannot show that developers lack unrelated administrative powers.
D:Does not show access scope. Login success alone cannot show that developers lack unrelated administrative powers.
Question 176
A platform team has been granting project Editor permissions to application developers who only need to deploy one service. Which operational action best maintains permission drift management after launch?
- Preserve old temporary role grants indefinitely
- Review only firewall logs to find unused IAM permissions
- Audit access once at first deployment only
- Review role usage and group membership periodically and after team changes
Explanation
Correct Answer
D:Least privilege decays when temporary or unused grants remain indefinitely
Incorrect Answers
A:Allows privilege accumulation. Least privilege decays when temporary or unused grants remain indefinitely.
B:Network telemetry is not authorization review. Least privilege decays when temporary or unused grants remain indefinitely.
C:Team membership will change. Least privilege decays when temporary or unused grants remain indefinitely.
Question 177
An external CI system builds containers for a Google Cloud project and currently stores long-lived service-account JSON keys. Which architecture most directly meets the need for keyless external workload authentication?
- Store service-account JSON keys in encrypted CI variables for years
- Share one service account private key across multiple CI vendors
- Use a production user password in the build job
- Use Workload Identity Federation to exchange approved external assertions for scoped credentials
Explanation
Correct Answer
D:Federation avoids persistent downloadable service-account private keys
Incorrect Answers
A:Long-lived exportable keys still create theft and rotation risks. Federation avoids persistent downloadable service-account private keys.
B:Destroys trust separation. Federation avoids persistent downloadable service-account private keys.
C:Not a scoped workload identity. Federation avoids persistent downloadable service-account private keys.
Question 178
An external CI system builds containers for a Google Cloud project and currently stores long-lived service-account JSON keys. Which implementation choice most directly delivers restricting external trust?
- Accept external assertions from any repository under the same provider issuer
- Map arbitrary user-supplied CI variables to privileged workload identity claims
- Let all external jobs impersonate the same high-privilege administrator
- Limit the workload identity provider to exact issuer, audience, and repository or job claims
Explanation
Correct Answer
D:An unscoped trust configuration could accept credentials from unrelated external jobs
Incorrect Answers
A:Trusts unapproved builds. An unscoped trust configuration could accept credentials from unrelated external jobs.
B:Claims must be issuer-authenticated. An unscoped trust configuration could accept credentials from unrelated external jobs.
C:Unnecessary privilege across repositories. An unscoped trust configuration could accept credentials from unrelated external jobs.
Question 179
An external CI system builds containers for a Google Cloud project and currently stores long-lived service-account JSON keys. Which validation would best establish validation of issuer boundary?
- Present a valid authorized build assertion and a forged or unapproved repository claim
- Test only a legitimate repository build
- Check that a pool object exists without a token exchange
- Verify the Google API is reachable without using the external identity
Explanation
Correct Answer
A:Negative token-exchange tests prove the trust condition rejects wrong external identities
Incorrect Answers
B:Cannot prove unapproved assertions are denied. Negative token-exchange tests prove the trust condition rejects wrong external identities.
C:Does not verify attributes. Negative token-exchange tests prove the trust condition rejects wrong external identities.
D:Not an identity-boundary test. Negative token-exchange tests prove the trust condition rejects wrong external identities.
Question 180
An external CI system builds containers for a Google Cloud project and currently stores long-lived service-account JSON keys. Which operational action best maintains rotating third-party trust safely after launch?
- Keep retired provider mappings active indefinitely
- Rotate only CI usernames while leaving federation policy unchanged
- Monitor only the artifact upload size
- Review federation claim mappings and revoke unneeded pools/providers after workflow changes
Explanation
Correct Answer
D:External CI trust must track the lifecycle of the authorized build system
Incorrect Answers
A:Old jobs may retain trust. External CI trust must track the lifecycle of the authorized build system.
B:Does not remove stale trust attributes. External CI trust must track the lifecycle of the authorized build system.
C:Misses credential exchange anomalies. External CI trust must track the lifecycle of the authorized build system.
Google Cloud Certified - Professional Cloud Architect (Professional Cloud Architect) Practice Tests