Google Cloud Certified Professional Cloud Developer (PCD) practice questions

Exam overview PCD

Sample questions 9 of 84 shown

1. Q1 Designing highly scalable, available, and reliable cloud-native applications

   You are developing a scalable web application for internal users. Your organization uses Google Workspace. You need to set up authentication to the application for the users, and then deploy the application on Google Cloud. You plan to use cloud-native features, and you want to minimize infrastructure management effort. What should you do? (Choose two.)

   1. Create a Compute Engine VM, configure a web server, and deploy the application in a VPC.
   2. Containerize the application, and deploy it as a Cloud Run service.
   3. Configure Cloud SQL database with a table containing the users and password hashes. Add an authentication screen to ensure that only internal users can access the application.
   4. Configure Identity Aware Proxy, and grant the roles/iap.httpsResourceAccessor IAM role to the users that need to access the application.

   BD. Cloud Run (B) provides serverless container execution with zero infrastructure management for Google Workspace users. IAP (D) secures internal apps using Google Workspace identity without managing auth code. Compute Engine VM with VPC (A) adds infrastructure overhead. Custom auth with Cloud SQL (C) duplicates Google Workspace identity management.
2. Q2 Designing highly scalable, available, and reliable cloud-native applications

   Your team currently uses Bigtable as their database backend. In your application's app profile, you notice that the connection to the Bigtable cluster is specified as single-cluster routing, and the cluster’s connection logic is configured to conduct manual failover when the cluster is unavailable. You want to optimize the application code to have more efficient and highly available Bigtable connectivity. What should you do?

   1. Set up Memcached so that queries hit the cache layer first and automatically get data from Bigtable in the event of a cache miss.
   2. Increase the Bigtable client’s connection pool size.
   3. Configure a Dataflow template, and use a Beam connector to stream data changes.
   4. Configure the app profile to use multi-cluster routing.

   D. Memorystore for Redis caches frequent Bigtable reads with sub-millisecond latency, reducing load on the database. Short-lived sessions (A) don't provide data caching. Batch processing (B) reduces frequency but not per-query latency. Increasing Bigtable nodes (C) costs more and doesn't match caching performance.
3. Q3 Designing highly scalable, available, and reliable cloud-native applications

   You are a developer that works for a local concert venue. Customers use your company’s website to purchase tickets for events. You need to provide customers with immediate confirmation when a selected seat has been reserved. How should you design the ticket ordering process?

   1. Add the seat reservation to a Cloud Tasks queue, which triggers Workflows to process the seat reservation.
   2. Publish the seat reservation to a Pub/Sub topic. Configure the backend service to use Eventarc to process the seat reservation on GKE.
   3. Upload the seat reservation to a Cloud Storage bucket, which triggers an event to a Cloud Run service that processes the orders.
   4. Submit the seat reservation in an HTTP POST request to an Application Load Balancer. Configure the Application Load Balancer to distribute the request to a Compute Engine managed instance group that processes the reservation.

   D. Serverless NEGs (network endpoint groups) enable Cloud Run services to be backends for Cloud Load Balancing with IAP. Configuring Cloud DNS (A) only resolves names. Direct VPC connection (B) bypasses load balancing. Internal HTTP LB (C) doesn't integrate with IAP for authentication.
4. Q4 Designing highly scalable, available, and reliable cloud-native applications

   You have an application running on a GKE cluster. Your application has a stateless web frontend, and has a high-availability requirement. Your cluster is set to automatically upgrade, and some of your nodes need to be drained. You need to ensure that the application has a serving capacity of 10% of the Pods prior to the drain. What should you do?

   1. Configure a Vertical Pod Autoscaler (VPA) to increase the memory and CPU by 10% and set the updateMode to Auto.
   2. Configure the Pod replica count to be 10% more than the current replica count.
   3. Configure a Pod Disruption Budget (PDB) value to have a minAvailable value of 10%.
   4. Configure the Horizontal Pod Autoscaler (HPA) maxReplicas value to 10% more than the current replica count.

   B. Cloud CDN caches static content at Google's edge locations, reducing latency for global users and offloading origin servers. Multi-cluster Ingress (A) adds complexity. Cloud NAT (C) provides outbound internet. External HTTP(S) LB alone (D) distributes but doesn't cache.
5. Q5 Designing highly scalable, available, and reliable cloud-native applications

   Your company manages an application that captures stock data in an internal database. You need to create an API that provides real-time stock data to users. You want to return stock data to users as quickly as possible, and you want your solution to be highly scalable. What should you do?

   1. Create a BigQuery dataset and table to act as the internal database. Query the table when user requests are received.
   2. Create a Memorystore for Redis instance to store all stock market data. Query this database when user requests are received.
   3. Create a Bigtable instance. Query the table when user requests are received. Configure a Pub/Sub topic to queue user requests that your API will respond to.
   4. Create a Memorystore for Redis instance, and use this database to store the most accessed stock data. Query this instance first when user requests are received, and fall back to the internal database.

   D. Cloud Endpoints provides API management with authentication, monitoring, and quotas for HTTP APIs. Apigee (A) is for enterprise API platforms. Cloud Tasks (B) is for async task distribution. Cloud Scheduler (C) triggers jobs on schedule, not API management.
6. Q6 Designing highly scalable, available, and reliable cloud-native applications

   You are designing a microservices architecture for a new application that will be deployed on Cloud Run. The application requires high-throughput communication between the internal microservices. You want to use the most effective, lowest latency communication protocol for this application. What should you do?

   1. Configure the Cloud Run service to use HTTP/2. Implement gRPC for communication between the microservices. Use streaming gRPCs when a large amount of data has to be sent.
   2. Implement the microservices with the REST API communication protocol. Use Apigee with rate-limiting to provide the best QoS for high-priority services.
   3. Use SOAP to build the microservices API, and use XML as the data format for communication across the microservices. Define SOAP data contracts for each microservice.
   4. Use HTTP REST to communicate across the microservices. Implement pagination and add indexing to your database.

   A. Cloud Run with per-request billing and scale-to-zero is ideal for processing variable image workloads. GKE (B) has minimum node costs. App Engine (C) has flexible scaling but different pricing. Compute Engine (D) is always-on IaaS.
7. Q7 Designing highly scalable, available, and reliable cloud-native applications

   You are developing an application that will allow users to read and post comments on news articles. You want to configure your application to store and display user-submitted comments using Firestore. How should you design the schema to support an unknown number of comments and articles?

   1. Store each comment in a subcollection of the article.
   2. Add each comment to an array property on the article.
   3. Store each comment in a document, and add the comment's key to an array property on the article.
   4. Store each comment in a document, and add the comment's key to an array property on the user profile.

   A. Firestore provides real-time sync and offline support for mobile/web apps, ideal for comment systems. Cloud SQL (B) lacks real-time sync. Bigtable (C) is for analytical workloads at scale. Cloud Storage (D) is for objects, not real-time sync.
8. Q8 Designing highly scalable, available, and reliable cloud-native applications

   You are a developer working with the CI/CD team to troubleshoot a new feature that your team introduced. The CI/CD team used HashiCorp Packer to create a new Compute Engine image from your development branch. The image was successfully built, but is not booting up. You need to investigate the issue with the CI/ CD team. What should you do?

   1. Create a new feature branch, and ask the build team to rebuild the image.
   2. Shut down the deployed virtual machine, export the disk, and then mount the disk locally to access the boot logs.
   3. Install Packer locally, build the Compute Engine image locally, and then run it in your personal Google Cloud project.
   4. Check Compute Engine OS logs using the serial port, and check the Cloud Logging logs to confirm access to the serial port.

   D. Cloud Build with Terraform builder manages infrastructure as code in CI/CD pipelines. Helm (A) manages K8s apps. Separate CI/CD (B) fragments pipeline. Manual steps (C) are error-prone.
9. Q9 Designing highly scalable, available, and reliable cloud-native applications

   You manage an application that runs in a Compute Engine instance. You also have multiple backend services executing in stand-alone Docker containers running in Compute Engine instances. The Compute Engine instances supporting the backend services are scaled by managed instance groups in multiple regions. You want your calling application to be loosely coupled. You need to be able to invoke distinct service implementations that are chosen based on the value of an HTTP header found in the request. Which Google Cloud feature should you use to invoke the backend services?

   1. Traffic Director
   2. Service Directory
   3. Anthos Service Mesh
   4. Internal HTTP(S) Load Balancing

   A. Internal HTTP(S) Load Balancer provides private communication between Compute Engine and backend services within the VPC. External LB (B) exposes services publicly. Cloud NAT (C) provides outbound internet. VPN (D) connects on-premises.

Answer all 84 questions with instant feedback →

Reader reviews 208 ratings

• The PCD practice exam was crucial to my success. The domains map perfectly to the official exam blueprint.

  ★★★★★★ Jennifer F.

• I was really impressed by the quality of the PCD questions. No typos, no vague wording — they clearly know the material.

  ★★★★★★ Stella P.

• Recommended this to my entire study group for PCD. Four out of four passed. That says something.

  ★★★★★★ Audrey K.

• Detailed, organized, and accurate. Exactly what you want in PCD prep material. The explanations deserve special mention.

  ★★★★★★ Gabriel L.

• Only used this for a week before my PCD exam and it was enough. Questions are very close to the real deal.

  ★★★★★★ Katherine W.

• Passed PCD with 912/1000. The practice questions cover the exam objectives thoroughly and the explanations are clear.

  ★★★★★ Ella M.

Frequently asked questions

More Google exams