Try 10 Real Questions From Google Professional Cloud Architect
These are unedited questions from the full 550-question set, pulled from across every exam domain. Click a question to reveal the answer and the full rationale for every option, exactly as it appears in the PDF.
A national grocery chain is moving its promotions platform to Google Cloud. Leadership has stated that the migration must not interrupt weekend trading, that infrastructure cost must fall by at least 20 percent in year one, and that the marketing team must be able to launch a campaign without filing a ticket. The architect is asked to translate these statements into design inputs. How should the architect classify them?
Show answer & full rationale
A (Correct): Availability and cost targets describe how well the system must behave and are therefore non-functional, while the ability for marketing to launch a campaign is a capability the system must provide and is therefore functional.
B: Being quantified does not make a requirement more valid; availability and self-service goals are equally binding and can also be made measurable.
C: Cost, availability and self-service targets directly shape platform and service selection and cannot be deferred outside the design.
D: The source of a requirement does not determine its type; classification depends on whether it describes a capability or a quality attribute.
An insurance company is preparing a business case for migrating a claims platform. The CFO wants a single figure that shows whether the migration is worth funding over three years. The architect has estimated migration labor, license buy-outs, run-rate cloud spend, and the value of decommissioned data center space. Which measure should the architect present?
Show answer & full rationale
A: Migration throughput is a delivery metric and says nothing about financial return.
B: Steady-state invoice ignores migration cost, retired data center cost and license changes, so it cannot show whether the investment pays back.
C: Throughput capacity is a technical characteristic, not a financial justification.
D (Correct): ROI expressed over the investment horizon captures both the spend and the realized benefit, which is what a funding decision requires.
An organization has two Dedicated Interconnect circuits terminating in different metropolitan areas and wants the highest available redundancy for hybrid connectivity, with automatic failover. Which configuration should be used?
Show answer & full rationale
A (Correct): Placing attachments in separate edge availability domains with BGP sessions on redundant Cloud Routers gives automatic failover if any single element fails.
B: Static routes do not fail over automatically and a single Cloud Router is a shared failure point.
C: A cold standby requires human intervention, which is the opposite of automatic failover.
D: Adding VPN over private circuits introduces encryption overhead and complexity without improving the failover behavior.
A new requirement means the team must allow a workload in a service project to reach an on-premises network over a Cloud Interconnect terminated in the host project's Shared VPC. What configuration is required?
Show answer & full rationale
A: Service project workloads sit on the Shared VPC, so a separate VPC and tunnel is unnecessary duplication.
B: This defeats the purpose of the private interconnect.
C (Correct): Service project workloads use the host project's network, so hybrid reachability depends on the host project's Cloud Router advertising those ranges.
D: Shared VPC attachment already provides the network; peering is a different and unnecessary mechanism here.
An organization wants environment-specific policies applied consistently: production must forbid external IP addresses on virtual machines, while a sandbox may allow them. Which design achieves this with least ongoing effort?
Show answer & full rationale
A: Firewall rules govern traffic and do not prevent an external address being assigned to an instance.
B: Blocking instance creation entirely is far broader than the requirement and prevents legitimate work.
C: Per-project application depends on someone remembering for every new project and drifts over time.
D (Correct): Policies inherit down the hierarchy, so applying the constraint at the folder level covers all current and future production projects automatically.
A workload's service account currently holds the Editor role on its project. The security team requires least privilege. What is the most appropriate approach?
Show answer & full rationale
A: This reproduces the same excessive access under a different name.
B (Correct): Basing the replacement on observed usage produces a role set that is minimal but sufficient, rather than a guess that breaks the workload.
C: Discovering missing permissions through production failures is disruptive and unsafe.
D: A time condition does not reduce the breadth of what the identity may do while it is active.
A team's deployments succeed in staging but fail intermittently in production. Investigation shows the two environments were built at different times by different people using the console. What process change addresses the root cause?
Show answer & full rationale
A: The evidence points to configuration divergence rather than capacity.
B: Manual comparison is slow, incomplete and drifts again immediately after each deployment.
C: More frequent deployment of a divergent environment produces more failures rather than fewer.
D (Correct): A single parameterized definition removes the configuration divergence that causes environment-specific failures.
An organization wants to reduce the time between a change being merged and reaching production, without increasing the change failure rate. Which practice most directly achieves both?
Show answer & full rationale
A: Larger batches increase both the lead time and the difficulty of diagnosing a failure.
B: An additional manual gate lengthens lead time without improving verification quality.
C: Centralizing deployment on one person creates a bottleneck and a single point of knowledge.
D (Correct): Small automated increments both shorten lead time and reduce failure rate, because each change is smaller and consistently verified.
A development team is about to deploy a new version of a service that changes the database schema. The team asks how to sequence the deployment so the service stays available. What should the architect advise?
Show answer & full rationale
A: The new application version cannot function without the schema it depends on.
B: Planned downtime is unnecessary when a compatible sequence exists.
C (Correct): The expand-then-contract sequence keeps the schema compatible with both the old and new application versions throughout the rollout.
D: A simultaneous change leaves a window in which some instances run against an incompatible schema.
An architect is applying the operational excellence pillar of the Google Cloud Well-Architected Framework to a platform team. Which practice most reflects the pillar's principles?
Show answer & full rationale
A: Utilisation is a capacity signal and does not reflect operational readiness or service health.
B (Correct): The pillar centers on readiness, automation of toil and learning from operational events, which this practice combines.
C: Manual gates add friction and delay without improving reliability, which the pillar treats as toil.
D: Concentrating knowledge in individuals creates key-person risk rather than operational maturity.
Google Professional Cloud Architect — $29.99, instant PDF download with full rationale on every answer.
View the practice questions →