Back to practice examsAdvanced

Google Cloud Developer Professional

Tests ability to build scalable applications, integrate Google Cloud services, and optimize app performance

Time limit

2h

Questions

65

Passing score

75%

Difficulty

Advanced

Category

Google Cloud Developer Professional

Sample questions

A few questions from this practice exam, with the correct answer and why it is correct.

  1. 1.A team packages a stateless HTTP API as a container. Traffic is bursty, drops to zero at night, and the team does not want to manage nodes, clusters, or autoscaling policies. Which platform best fits these requirements?

    • ACloud RunCorrect answer
    • BGoogle Kubernetes Engine Standard
    • CCompute Engine managed instance groups
    • DApp Engine flexible environment

    Why

    Cloud Run runs stateless containers, scales automatically with request volume, scales to zero when idle, and requires no cluster or node management. GKE Standard requires managing a cluster and node pools. Compute Engine requires managing virtual machines and instance groups. App Engine flexible environment keeps at least one instance running and does not scale to zero.

  2. 2.An application needs sidecar containers, per-node daemons for log collection, custom pod scheduling based on node affinity, and StatefulSets with persistent volumes. Which platform should the team choose?

    • AGoogle Kubernetes EngineCorrect answer
    • BCloud Run services
    • CCloud Run jobs
    • DCloud Functions

    Why

    Google Kubernetes Engine exposes the full Kubernetes API, including DaemonSets, StatefulSets, affinity rules, and sidecars. Cloud Run services run a single container group with a much narrower model and no DaemonSets or node scheduling control. Cloud Functions runs individual functions. Cloud Run jobs execute finite tasks rather than long-lived stateful workloads.

  3. 3.A web application serves users in Europe, North America, and Asia from a single anycast IP address, and must route each user to the closest healthy backend while terminating TLS at the edge. Which load balancer should be used?

    • AGlobal external Application Load BalancerCorrect answer
    • BRegional external Application Load Balancer
    • CInternal Application Load Balancer
    • DExternal passthrough Network Load Balancer

    Why

    The global external Application Load Balancer provides a single anycast IP, terminates TLS at Google's edge, and routes each request to the nearest healthy backend. A regional external Application Load Balancer serves one region only. An internal load balancer is not reachable from the internet. A passthrough Network Load Balancer operates at layer 4 and does not terminate TLS or make HTTP routing decisions.

  4. 4.A legacy shopping application keeps the cart in the memory of the instance that first served the user, and users lose their carts when requests land on different backends. Without rewriting the application immediately, what should the developer configure on the load balancer?

    • ASession affinity on the backend serviceCorrect answer
    • BA larger number of backend instances
    • CA shorter health check interval
    • DCloud CDN in front of the backend service

    Why

    Session affinity on the backend service directs requests from the same client to the same backend instance, which keeps the in-memory cart reachable. Increasing the instance count makes the problem more likely, not less. A shorter health check interval affects failure detection. Cloud CDN caches responses and would serve stale cart data rather than fix affinity.

  5. 5.A product catalogue service reads the same rows from Cloud SQL thousands of times per second, and read latency is dominating the response time. The data changes a few times per hour. What is the most effective change?

    • AAdd a Memorystore for Redis cache in front of the databaseCorrect answer
    • BAdd Cloud SQL read replicas in the same region
    • CIncrease the Cloud SQL instance machine size
    • DMigrate the catalogue to Bigtable

    Why

    Putting a Memorystore for Redis cache in front of the database serves the repeated reads from memory in sub-millisecond time and removes most of the load from Cloud SQL, which suits data that changes rarely. Adding read replicas helps throughput but still pays database latency on every read. A larger instance is more expensive and does not remove the round trip. Moving to Bigtable is a disproportionate rewrite for a small catalogue.

Practice all 65 questions

Frequently asked questions

Is the Google Cloud Developer Professional practice exam free?

You can take it with the free attempts included in your account. After those run out, this exam requires a premium plan, while foundational-level exams stay free and unlimited.

How many questions does it have and how long does it take?

65 questions with a 120-minute time limit, so you practise under the same time pressure as the real exam.

What score do I need to pass?

You need 75% to pass this practice exam. Your score and a per-question review are shown as soon as you finish.

Do I need an account?

Yes, a free account. It is what lets us save your attempt, score it and keep your history.

Can I retake it?

Yes. Each attempt uses one of your free attempts, and a premium plan makes them unlimited.

Are these the real questions from the official exam?

No. NaHero does not reproduce questions from the official exam. These are practice questions written to match its format, topics and difficulty.

Google Cloud Developer Professional