Practice questions · IT & Cloud
Google Cloud Professional Cloud Architect: Practice Questions
Original concept-check questions for the Google Professional Cloud Architect (PCA). They test the design judgement the exam rewards - designing, provisioning, security, reliability and optimisation on Google Cloud - with every answer explained, including why the others are wrong. Filter by area or difficulty. These are concept checks, not real exam questions.
Answered 0 · Correct 0
-
The Google Professional Cloud Architect exam is best described as testing your ability to:
Correct answer: C. The PCA validates designing GCP solutions that satisfy business and technical requirements. Writing application code in Go is a developer skill, managing a help desk is a support role, and administering a single VM is a narrow operations task rather than solution design. -
For a stateless web service that should scale to zero when idle, the most fitting GCP service is:
Correct answer: D. Cloud Run runs stateless serverless containers and scales to zero when idle. A constantly running Compute Engine VM keeps paying for idle capacity, Bigtable is a NoSQL database, and Cloud Interconnect is a hybrid networking link, none of which host a scale-to-zero web service. -
You need managed Kubernetes for containerised microservices. The natural choice is:
Correct answer: C. GKE is Google's managed Kubernetes service for running containers. App Engine is a PaaS that abstracts away Kubernetes, Cloud SQL is a managed relational database, and Cloud KMS manages encryption keys, so none of them give you Kubernetes. -
A globally distributed, strongly consistent relational database that scales horizontally points to:
Correct answer: C. Spanner is globally distributed, strongly consistent and horizontally scalable relational. Cloud SQL is a regional managed relational database that does not scale horizontally across regions, Firestore is a NoSQL document database, and Cloud Storage holds objects, not relational tables. -
For serverless SQL analytics over very large datasets, use:
Correct answer: B. BigQuery is a serverless data warehouse for SQL analytics over very large datasets. Compute Engine is raw VMs you must manage, Cloud CDN caches content at the edge, and Cloud VPN connects networks, none of which run analytics queries. -
A requirement says 'lowest cost while meeting the SLA'. The best architecture is the one that:
Correct answer: A. The constraint is met by the design that satisfies the SLA at the lowest cost. Using the most premium services, running multi-region active-active by default, or adding maximum redundancy all push cost above what the SLA requires. -
In Google Cloud, a VPC network is:
Correct answer: A. A GCP VPC is a global resource that spans regions within a project, with regional subnets inside it. It is not regional-only, not a database, and not a monitoring dashboard. -
To serve a global user base behind a single anycast IP with health-checked backends, use:
Correct answer: C. Global Cloud Load Balancing offers a single anycast IP and routes traffic to healthy backends worldwide. A single VM's public IP has no load balancing or health checks, while Cloud KMS manages keys and Bigtable is a database. -
Object data such as images and backups belongs in:
Correct answer: A. Cloud Storage is built for unstructured object data like images and backups. Cloud SQL stores relational tables, Spanner is a global relational database, and Pub/Sub is a messaging service, so none suit object storage. -
Moving rarely accessed Cloud Storage data to the 'archive' storage class primarily:
Correct answer: C. The archive class is the cheapest to store but has higher retrieval cost and latency, so it suits rarely accessed data. It does not speed up frequent access, turn objects into a database, or add a second layer of encryption. -
For dedicated, private connectivity between on-premises and Google Cloud, design with:
Correct answer: B. Cloud Interconnect provides dedicated, private connectivity between on-premises and Google Cloud. Cloud CDN caches web content, BigQuery is an analytics warehouse, and Firestore is a NoSQL database, none of which is a hybrid network link. -
Google Cloud IAM should be designed around the principle of:
Correct answer: C. Least privilege grants each identity only the access it needs. Sharing one account removes accountability, granting everyone Owner is dangerously broad, and disabling logging destroys your audit trail. -
A service account is used to:
Correct answer: B. A service account is a non-human identity that lets applications and VMs authenticate to GCP services. It does not store object data (that is Cloud Storage), replace the billing account, or define a subnet (a VPC concept). -
To enforce policy and access across many projects at once, use the:
Correct answer: A. The resource hierarchy (organisation, folders, projects) lets you apply IAM and policies that inherit down to many projects at once. Cloud CDN, a single VM, and a storage bucket are individual resources, not organisation-wide policy scopes. -
Cloud KMS is used to:
Correct answer: B. Cloud KMS creates and manages cryptographic keys. Running containers is GKE or Cloud Run, streaming events is Pub/Sub, and balancing load is Cloud Load Balancing. -
A regulatory requirement to keep data within a specific region is best addressed by:
Correct answer: A. Data residency is met by selecting resource locations and regions that comply with the requirement. Disabling IAM weakens security, ignoring location fails the rule outright, and the archive storage class only affects cost, not region. -
For metrics, dashboards and alerting on GCP systems, use:
Correct answer: B. Cloud Monitoring collects metrics and provides dashboards and alerting. Spanner is a database, Cloud Storage holds objects, and Cloud Interconnect is a network link, so none of them deliver observability. -
To centralise, search and analyse logs across GCP, use:
Correct answer: B. Cloud Logging centralises log data so you can search and analyse it. Cloud VPN connects networks, Cloud CDN caches content, and Bigtable is a NoSQL database, none of which manage logs. -
Deploying across multiple zones in a region primarily protects against:
Correct answer: D. Spreading instances across zones keeps the service running if a single zone (datacentre) fails. Billing errors, exam fees, and slow code are not failure modes that multi-zone deployment addresses. -
To meet a near-zero RPO for a critical dataset, the design should favour:
Correct answer: D. A near-zero recovery point objective requires continuous or synchronous replication so almost no data is lost on failure. Infrequent manual exports, backups on the same disk, and turning off redundancy all leave large windows of unrecoverable data. -
An SLO (service level objective) in a reliability design is:
Correct answer: D. An SLO is a target level of reliability the team commits to, such as 99.9% availability. It is not a network protocol, a billing tier, or a type of VM. -
To reduce cost on predictable, steady Compute Engine workloads, consider:
Correct answer: A. Committed use discounts lower the price of steady, predictable Compute Engine workloads in exchange for a usage commitment. Disabling autoscaling and deleting monitoring do not cut compute cost, and paying on-demand list price is the most expensive option. -
Autoscaling a managed instance group helps optimise:
Correct answer: B. Autoscaling adds and removes instances to match capacity to demand, optimising both cost and performance. Key management (KMS), DNS resolution (Cloud DNS), and an exam pass mark are unrelated to instance group scaling. -
Adding Cloud CDN in front of static content primarily improves:
Correct answer: B. Cloud CDN caches static content at edge locations close to users, cutting latency. It does not change IAM policy, database write speed, or state locking. -
Right-sizing recommendations in Google Cloud help you:
Correct answer: A. Right-sizing matches machine types to actual usage so you stop paying for idle capacity. Increasing every VM to the largest size wastes money, while removing backups and disabling logging sacrifice safety without right-sizing anything. -
When two designs both meet the requirements, the PCA-preferred answer usually:
Correct answer: A. When two designs both meet the requirements, prefer the one that costs less and is simpler to operate. Using the most services or choosing the hardest design to maintain adds needless complexity, and ignoring the requirements disqualifies a design. -
The published case studies in the PCA exam are best used by:
Correct answer: C. Several PCA questions reference the published case studies, so reading them in advance and working out each company's architecture pays off. Memorising only company names, assuming they will not be tested, or ignoring them until exam day all waste that preparation. -
When a case study states a strict compliance requirement, your design must:
Correct answer: D. A stated compliance requirement is a binding constraint your design must satisfy. Picking the newest service regardless, optimising only for cost, or skipping the question all leave the requirement unmet. -
For event-driven integration between decoupled GCP services, a natural choice is:
Correct answer: B. Pub/Sub is Google's messaging service for decoupled, event-driven integration between services. Spanner is a relational database, Cloud CDN caches content, and Cloud KMS manages keys, none of which carry events between services. -
A 'lift-and-shift' migration to Google Cloud most directly maps to:
Correct answer: A. Lift-and-shift means rehosting existing workloads with minimal change, typically on Compute Engine VMs. Deleting the workloads, rewriting everything as Cloud Functions first, or moving them into BigQuery all involve far more change than rehosting.
Practice questions FAQ
- Are these real PCA exam questions?
- No. These are original study questions written to test understanding. They are not real exam questions, exam dumps, or copied from any provider.
- How should I use these practice questions?
- Answer each one, read the explanation (including why the wrong options are wrong), and use the per-domain score below to focus your revision on weak areas. Revisit before exam day.
- How many questions should I do before the exam?
- Enough to score consistently across every domain, alongside full-length practice from official or reputable providers. Understanding why each answer is right matters more than raw volume.
- What score means I am ready?
- A good signal is consistently scoring around 80% or higher across all domains on questions you have not seen before, and being able to explain why the wrong options are wrong.
- Should I use exam dumps?
- No. Dumps (real or leaked questions) breach provider policy, can void your certification, and do not build the understanding the exam actually tests.