Google Cloud Professional Cloud Architect (PCA) Readiness Test

10 questions, balanced across the exam's topics. Answer each one, then submit for your practice readiness estimate and per-topic breakdown. This is a practice estimate — not a prediction or guarantee of your official exam result.

After your result, you can start studying right away: the first questions of the full practice bank are free, with no card required.

Question 1

A hospital imaging service currently restores from nightly backups in 8 hours with up to 24 hours of data loss. The board now requires RTO under 30 minutes and RPO under 5 minutes after a regional disaster, and approves significant spend. Which change in disaster recovery strategy is required?

Keep backup-and-restore but compress the backup files so restoration is faster Move to a pilot light design where data is restored from backup after the disaster is declared Move to a warm standby or active-active design in a second region with continuous replication, pre-provisioned capacity, and automated failover Increase backup frequency to every 12 hours and keep the same restore process

Question 2

Workloads in a VPC must call Google APIs without traffic traversing the public internet, and the compliance team also wants API access restricted to specific services with regional control. Which combination should the architect choose?

Cloud NAT with logging enabled on all subnets Public IP addresses on every VM protected by Cloud Armor Private Google Access / Private Service Connect endpoints for Google APIs, so requests stay on Google's network and can be governed by VPC and regional policies A site-to-site VPN to Google headquarters for API traffic

Question 3

A service has a 99.9% availability SLO over a rolling 30-day window. Error budget burn is accelerating after a release, and a burn rate alert fires for fast consumption over 1 hour and 6 hours. What is the purpose of these multi-window burn rate alerts?

To detect both rapid budget exhaustion needing immediate action and slower burns needing investigation before the budget is gone To automatically roll back any release without human review To replace the need for an SLO entirely To measure how much logging budget the service has consumed

Question 4

A project will cost $500,000 to deliver. It is expected to generate $150,000 per year in net benefit for five years. What is the simple ROI over the five years, and what does it mean?

ROI is 30%, meaning the project returns its cost in the first year ROI is 250%, meaning benefits are two and a half times the cost every year ROI is 50%, meaning total net gain of $250,000 on a $500,000 investment over five years ROI cannot be estimated until the project is complete

Question 5

An architect needs to capture an entire VM — boot disk plus all attached data disks, metadata, and permissions — so an identical instance can be recreated later, for example for a backup or a clone of a complex stateful server. Which artifact is designed for this?

A Persistent Disk snapshot of the boot disk A machine image, which captures the configuration, metadata, permissions, and contents of all disks of an instance A custom image of the boot disk An instance template

Question 6

A development team asks the architect whether to optimize its new service for the lowest possible cost or the lowest possible latency. Leadership has stated the service supports a revenue-critical checkout flow with a strict latency budget in the service-level objectives. How should the architect advise the team?

Choose the cheapest design, because cost is always the architect's first responsibility Choose the lowest-latency design regardless of how far it exceeds the budget Defer the decision until after launch when real traffic data is available Prioritize meeting the latency SLO for the checkout flow, then reduce cost within that constraint, and document the trade-off and its cost impact for leadership

Question 7

An enterprise wants name resolution for internal services (for example, db.internal.example.com) to work only for clients inside its VPCs, while the same domain also has public records for internet clients. Which Cloud DNS design should the architect implement?

Create a Cloud DNS private zone for internal records and a separate public zone for external records Create a Cloud DNS peering zone to hide the records from the internet Configure a DNS response policy zone for the internal names Store the records in instance metadata so only VMs can see them

Question 8

For incident investigation, a CISO requires audit logs from every project in the organization — including projects created in the future — to be retained centrally for at least one year in storage the project teams cannot alter, and to be queryable for investigations. What should the architect implement?

A log sink in each existing project exporting to its own Cloud Logging bucket with default retention Leave logs in each project's _Default bucket and rely on its 30-day retention A folder-level export to a Cloud Storage bucket in each project, reviewed manually An organization-level aggregated sink with includeChildren routing the required audit logs to a BigQuery dataset in a dedicated, tightly controlled logging project with retention set for at least a year

Question 9

A financial services company must protect Compute Engine and Cloud SQL backups against accidental deletion by a project administrator and keep a copy outside the primary region for regional disaster recovery. Which design best meets both requirements?

Store snapshots only in the primary region and rely on IAM deny policies Use local persistent disk snapshots with a 7-day retention rule Use Backup and DR Service with a backup vault in a second region and set a minimum enforced retention period on the vault Copy backup exports to a single-zone Cloud Storage bucket in the primary region

Question 10

A GKE batch analytics workload can tolerate interruption, runs for a few hours nightly, and dominates cluster cost. Interactive services in the same cluster cannot tolerate interruption. What is the best cost optimization?

Run the batch workload on a dedicated node pool using Spot VMs, keeping interactive services on standard nodes Move the interactive services to Spot VMs to share the discount Run the batch workload continuously so it never restarts Give the batch pods the highest priority class in the cluster permanently

Unanswered questions count as incorrect. Your answers are graded on the server; no answer key is sent to your browser.

← Back to the Google Cloud Professional Cloud Architect (PCA) page