Google Cloud Professional Cloud Architect Practice Test

Professional Cloud Architect practice test: what's covered

• Professional Cloud Architect

Free Professional Cloud Architect sample questions

1. Mountkirk Games is launching a multiplayer game that must serve players across North America, Europe, and Asia with a single globally consistent player-profile database and the lowest possible read latency. Which managed database satisfies the strong-consistency and global-scale requirement?

   • A. Cloud Spanner configured with a multi-region instance
   • B. Cloud SQL for PostgreSQL with a regional HA configuration
   • C. Firestore in Datastore mode within a single region
   • D. Cloud Bigtable cluster replicated within one region
   Show answer

   Correct: A. Cloud Spanner multi-region delivers horizontally scalable SQL with external (strong) consistency across continents and up to 99.999% availability, which is exactly what a global game profile store needs. Cloud SQL is regional and cannot offer globally consistent low-latency reads. Firestore in a single region does not span continents, and Bigtable is wide-column NoSQL without relational SQL or external consistency, so it is incorrect for a globally consistent relational profile store.

2. Dress4Win wants to migrate to Google Cloud while reducing operational toil and aligning spend with measurable business outcomes such as faster release cycles and predictable costs. Which design principles best support those stated business goals? (Choose TWO)

   • A. Prefer managed and serverless services to shift undifferentiated operations to Google
   • B. Adopt committed use discounts for steady-state baseline workloads
   • C. Run every workload on the largest available n2 instances to avoid resizing
   • D. Disable autoscaling so capacity stays fixed and easy to reason about
   • E. Manually patch each VM weekly to retain full control of the OS
   Show answer

   Correct: A and B. Using managed/serverless services reduces operational toil and speeds delivery, and committed use discounts make steady-state spend predictable and lower, both of which map directly to the business goals. Oversizing on the largest instances wastes money and is not justified by the goals. Disabling autoscaling removes elasticity and raises cost, whereas weekly manual patching increases toil instead of reducing it, so those options are incorrect.

3. An auditor asks how your business requirement of a 15-minute Recovery Time Objective and near-zero Recovery Point Objective translates into architecture decisions. Which mapping correctly reflects how RTO and RPO drive the design?

   • A. RTO mainly drives standby compute readiness while RPO drives data replication frequency
   • B. RTO and RPO both only affect how often you run backups, not replication
   • C. RPO drives how many regions host compute while RTO drives encryption choices
   • D. RTO and RPO are interchangeable and require identical mitigation
   Show answer

   Correct: A. A tight RTO requires warmer standby capacity (active-active or fast failover) so service is restored quickly, while a near-zero RPO requires synchronous or very frequent data replication to avoid data loss. Backups alone cannot meet a 15-minute RTO, so the backup-only option is incorrect. RPO does not govern compute region count and RTO does not govern encryption, and the two metrics are not interchangeable because they measure different things (downtime versus data loss).

4. A team wants to deploy a stateless containerized HTTP API that scales to zero when idle, requires no node or cluster management, and bills only for requests served. Which compute platform fits this requirement most directly?

   • A. Cloud Run fully managed service
   • B. Compute Engine managed instance group
   • C. GKE Standard cluster with a node pool
   • D. App Engine flexible environment
   Show answer

   Correct: A. Cloud Run runs stateless containers, scales to zero, requires no infrastructure management, and bills per request and CPU/memory used, matching all stated needs. A Compute Engine MIG and a GKE Standard cluster both require provisioning and managing infrastructure and do not scale to zero by default. App Engine flexible runs on VMs that do not scale to zero and is heavier, so it is not the best fit for the request-billed, scale-to-zero requirement.

5. A platform team needs Kubernetes but wants Google to manage nodes, apply hardened defaults, and bill based on the resources their pods actually request rather than per provisioned node. Which GKE mode should they choose?

   • A. GKE Autopilot mode
   • B. GKE Standard mode with custom node pools
   • C. Self-managed Kubernetes on Compute Engine
   • D. GKE Standard mode with node auto-provisioning only
   Show answer

   Correct: A. GKE Autopilot manages nodes for you, applies security-hardened defaults, and bills per pod resource request, exactly matching the requirement. GKE Standard, even with custom node pools or node auto-provisioning, still bills per provisioned node and leaves node management responsibilities with the team. Self-managed Kubernetes on Compute Engine pushes the most operational burden onto the team, so it is incorrect for a hands-off, pod-billed model.

Last updated: 2026-10-19

Professional Cloud Architect practice test FAQ

Related practice tests

• Google Cloud ACE practice test
• Google Professional Cloud DevOps Engineer practice test
• Google Professional Cloud Network Engineer practice test
• Google Cloud PCSE practice test
• Google Cloud Professional Data Engineer practice test
• Google Cloud GenAI-Leader practice test