Cloud Digital Leader Google Cloud products, services, and solutions Practice Test 1 — 15 Questions

Question 1

A data analytics team needs to run a one-time transformation on 10 TB of data stored in Cloud Storage, then load the results into BigQuery. The transformation is a custom Java application that reads files, processes them, and writes to a new location. Which service should they use to minimize operational overhead?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Cloud Dataflow with Apache Beam Java SDK. Cloud Dataflow with Apache Beam Java SDK is the appropriate service for running custom Java transformations on large datasets in a fully managed, serverless manner. It reads from Cloud Storage, processes data, and can load results to BigQuery without provisioning clusters. Other options like Cloud Functions have execution limits, GKE requires cluster management, and Dataproc Serverless is intended for Spark jobs, not arbitrary custom Java applications.

Answer

Google Kubernetes Engine (GKE) with a custom container

Answer

Dataproc Serverless with Spark job

Answer

Cloud Functions triggered by Cloud Storage events

Question 2

A company uses Cloud SQL for PostgreSQL and needs to run complex analytical queries on the same dataset without affecting the performance of the transactional database. What should they do?

Accepted Answer

Create read replicas of the Cloud SQL instance and run queries on the replicas. Creating read replicas of Cloud SQL allows running analytical queries on the replica without affecting the performance of the primary transactional database. Read replicas are fully managed and provide near-real-time data synchronization. Option A is incorrect because periodic exports to Cloud Storage and querying with BigQuery introduces latency and does not provide live querying on the same dataset. Option C is incorrect because simply upgrading the machine type does not isolate workloads; both transactional and analytical queries still compete for the same resources. Option D is incorrect because using BigQuery to directly query Cloud SQL via federated queries can still impact the primary instance's performance and is not optimized for complex analytical workloads; it is better suited for lightweight queries.

Answer

Schedule periodic exports to Cloud Storage and query with BigQuery

Answer

Upgrade the Cloud SQL instance to a higher machine type

Answer

Use BigQuery to directly query Cloud SQL via federated queries

Question 3

A company uses BigQuery for analytics and needs to enforce row-level security based on user department. Only users from the 'Sales' department should see rows where department = 'Sales'. Which BigQuery feature should they use?

Accepted Answer

Row-level access policies using a filter on the department column. Row-level access policies in BigQuery allow you to define filters that restrict rows based on user attributes, such as department. This enables row-level security by ensuring users only see rows they are authorized to access. Option A is incorrect because custom IAM roles control permissions at the resource level but not row-level filtering. Option B is incorrect because column-level security with classification tags hides entire columns, not rows. Option C is incorrect because authorized views share data across projects but do not inherently filter rows based on user identity.

Answer

Custom IAM roles with fine-grained permissions

Answer

Column-level security with classification tags

Answer

Authorized views with a WHERE clause

Question 4

A company wants to migrate a legacy monolithic application to Google Cloud with minimal code changes. The application runs on a specific OS version and uses a relational database. Which migration approach is most suitable?

Accepted Answer

Lift and shift to Compute Engine and Cloud SQL. 'lift and shift' (rehosting) moves the application to Compute Engine with minimal code changes, and Cloud SQL is a managed relational database service compatible with the existing database. Option A is wrong because rebuilding with serverless functions requires significant code changes and is not suitable for minimal changes. Option C is wrong because refactoring into microservices also requires substantial code changes. Option D is wrong because replatforming to Cloud Run may require containerization and platform adjustments.

Answer

Rebuild the application using serverless functions

Answer

Refactor the application into microservices

Answer

Replatform to use Cloud Run

Question 5

A company wants to implement a CI/CD pipeline for their Cloud Run services. Which THREE Google Cloud services should they use? (Choose 3)

Accepted Answer

Cloud Build. For a CI/CD pipeline for Cloud Run, you need: Cloud Build (to build the container image, run tests, and trigger the pipeline), Container Registry or Artifact Registry (to store the container image), and Cloud Deploy (to manage continuous delivery and rollout strategies like canary or blue/green deployments to Cloud Run). Cloud Operations Suite is for monitoring, and Cloud Deployment Manager is for infrastructure provisioning, so they are not part of the CI/CD pipeline.

Answer

Cloud Operations Suite

Answer

Cloud Deployment Manager

Question 6

A developer deploys a Cloud Function with the command shown:

gcloud functions deploy my-function --gen2 --region=us-central1 --runtime=python39 --trigger-http --allow-unauthenticated --timeout=300s

The function needs to process a file upload that typically takes 2 minutes. What is the most likely issue?

Accepted Answer

The memory allocation might be too low if the file is large. The command shown does not specify a memory allocation, so the Cloud Function defaults to 256 MB. If the file being uploaded is large, this low memory can cause the function to run out of memory and fail, even if the timeout is sufficient. Processing a file upload often requires loading the file into memory, making memory a critical resource.

Answer

The runtime python39 is not available

Answer

The function cannot be triggered by HTTP

Answer

The timeout is too short for processing a 2-minute upload

Question 7

A company is planning to migrate a legacy monolithic Linux application to Google Cloud. They want to minimize changes initially but have the flexibility to modernize later. Which three approaches should they consider?

Accepted Answer

Use Cloud Run for stateless containers. Cloud Run (A) allows deploying containerized monoliths with minimal code changes and provides a serverless environment that can later be refactored into microservices. Migrate for Anthos (C) enables moving existing VMs to GKE containers without changing the application, minimizing initial effort while allowing future modernization. Rehosting on Compute Engine (E) is a pure lift-and-shift with no code changes, offering the simplest path and later flexibility. In contrast, App Engine Flexible Environment (B) is a replatforming move that often requires application adaptation, and refactoring to microservices (D) is a major architectural change.

Answer

Replatform to App Engine Flexible Environment

Answer

Refactor to microservices on GKE

Question 8

A company is designing a disaster recovery plan for its critical application running on Google Kubernetes Engine (GKE) in us-central1. The application uses Cloud SQL (MySQL) as its database. Which TWO actions should the company take to achieve a Recovery Point Objective (RPO) of less than 5 minutes and a Recovery Time Objective (RTO) of less than 1 hour in the event of a regional outage? (Choose two.)

Accepted Answer

Configure Cloud SQL for cross-region replication using a read replica in a different region.. Cloud SQL cross-region replication using a read replica in a different region provides asynchronous replication with a typical lag of under 5 minutes, enabling an RPO of less than 5 minutes. In a regional outage, you can promote the read replica to a primary instance in the other region, achieving an RTO of under 1 hour by failing over the application's database connection. Additionally, deploying a GKE cluster in a different region with Multi-cluster Ingress allows traffic to be routed to the standby cluster, ensuring the application itself can fail over within the RTO. Together, these two actions provide both compute and database failover for the critical application.

Answer

Use Cloud SQL's automated backups with a backup schedule every 4 hours.

Answer

Store container images in a regional registry in the same region.

Answer

Use Cloud CDN to cache static content.

Question 9

A company runs a data processing pipeline on a single Compute Engine instance in us-west1-a. The instance reads data from Cloud Storage, processes it, and writes results back to Cloud Storage. The pipeline runs once per day and takes about 6 hours. Recently, the instance has been experiencing out-of-memory errors, causing the pipeline to fail. The operations team wants a cost-effective solution that can handle varying data volumes without manual intervention. They also want to ensure the pipeline completes within the daily window. What should they do?

Accepted Answer

Redesign the pipeline to run on Cloud Dataflow and use batch mode with autoscaling.. Cloud Dataflow's batch mode with autoscaling is designed for batch processing jobs like this one. It can handle varying data volumes by automatically scaling worker resources, ensuring the pipeline completes within the daily window. It is cost-effective because you only pay for the resources used. Option A is incorrect because managed instance groups with CPU autoscaling are not well-suited for a single long-running batch job; they are designed for scaling based on load but can lead to overprovisioning or insufficient resources. Option C is incorrect because streaming mode is for continuous data processing, not daily batch pipelines. Option D is incorrect because simply increasing memory does not provide elasticity for variable data volumes and may lead to increased costs without solving the scalability issue.

Answer

Use a managed instance group with autoscaling based on CPU utilization.

Answer

Redesign the pipeline to run on Cloud Dataflow and use streaming mode.

Answer

Increase the memory of the existing instance to a larger machine type.

Question 10

A financial services company is migrating a legacy monolithic application to Google Cloud. The application uses a SQL Server database and has compliance requirements to encrypt data at rest and in transit. The migration must minimize code changes. The application runs on Windows Server and currently uses Active Directory for authentication. The company wants to use Google Cloud's managed services where possible. Which approach best meets these requirements?

Accepted Answer

Lift and shift the application tier to Compute Engine on Windows VMs. Use Cloud SQL for SQL Server with encryption enabled, and configure Active Directory via Managed Microsoft AD.. Option B minimizes code changes by lifting and shifting the application to Compute Engine while using Cloud SQL for SQL Server, a managed database that supports encryption at rest and in transit. Managed Microsoft AD provides Active Directory authentication. The other options either keep the database on-premises, require containerization, or use unmanaged SQL Server on VMs with manual encryption and connectivity, so they do not best meet the managed services and compliance requirements.

Answer

Migrate the application to Compute Engine with SQL Server on Windows VMs. Use Cloud SQL for SQL Server but keep the database on-premises for compliance.

Answer

Containerize the application and deploy on GKE with SQL Server in a container. Use Cloud SQL for SQL Server with encryption.

Answer

Migrate the application to Compute Engine instances with SQL Server on Windows VMs. Use Cloud VPN for connectivity. Enable encryption on the VMs and use SSL for SQL connections.

Question 11

Given the Cloud Run service configuration above, what happens when a new revision is created after deploying a change to the container image?

Accepted Answer

The new revision receives all traffic automatically.. When you deploy a change to a Cloud Run service, a new revision is created and, by default, it receives 100% of traffic automatically. Traffic is not held on the old revision unless you explicitly use --no-traffic or configure a traffic split that excludes the new revision.

Answer

Traffic is split evenly between the old and new revision.

Answer

The service becomes unavailable until the new revision is ready.

Answer

Cloud Run automatically creates a new revision and routes traffic to it.

Answer

Traffic continues to go to the old revision until the traffic section is updated.

Question 12

A financial services company runs a multi-tier application on Google Kubernetes Engine (GKE). The application consists of a frontend service, a backend service, and a database on Cloud SQL. Recently, they noticed that the backend service experiences high latency during peak trading hours, causing the frontend to time out. The backend service is CPU-intensive and currently runs with a single replica. The team wants to reduce latency and ensure high availability without over-provisioning resources. They have enabled Horizontal Pod Autoscaling (HPA) based on CPU utilization with a target of 80% and default stabilization windows. However, during sudden traffic spikes, the HPA takes over 5 minutes to scale up because of the scale-up stabilization window and the time to trigger. The company cannot tolerate latency spikes during scaling. Which course of action should they take to minimize latency during traffic spikes?

Accepted Answer

Set the HPA target CPU utilization to 70% and remove the scale-up stabilization window (set to 0 seconds).. Reducing the HPA target CPU utilization to 70% provides a larger buffer before the threshold is crossed, allowing earlier scaling initiation. Removing the scale-up stabilization window (setting it to 0 seconds) eliminates the default 3-minute scale-up delay, enabling the HPA to react immediately to CPU spikes. This directly addresses the latency issue during sudden traffic spikes by reducing the time to add replicas.

Answer

Increase the HPA scale-up stabilization window to 10 minutes to prevent thrashing.

Answer

Enable session affinity on the backend service to keep users on the same pod.

Answer

Use a CronJob to proactively scale the backend service before expected peak hours.

Question 13

A data analytics team needs to analyze petabytes of structured data using SQL queries without managing any database infrastructure. Query results must return within seconds for most queries. Which Google Cloud service is designed for this use case?

Accepted Answer

BigQuery. BigQuery is a serverless, highly scalable data warehouse designed for analyzing petabytes of data using SQL without any infrastructure management. Its columnar storage and distributed query engine enable sub-second query performance on large datasets, making it ideal for this use case.

Answer

Cloud SQL

Answer

Cloud Bigtable

Answer

Cloud Spanner

Question 14

A developer wants to deploy a containerized web application without managing servers, clusters, or Kubernetes configuration. The application should automatically scale to zero when not in use and handle bursts of traffic. Which Google Cloud service is the best fit?

Accepted Answer

Cloud Run. Cloud Run is the best fit because it is a fully managed serverless container platform that automatically scales to zero when idle and scales up to handle traffic bursts, without requiring any server, cluster, or Kubernetes configuration. The developer simply deploys a container image, and Cloud Run handles all infrastructure management, including scaling and load balancing.

Answer

Google Kubernetes Engine (GKE)

Answer

Compute Engine

Answer

App Engine Standard

Question 15

A retail company wants to build a recommendation engine that suggests products to customers based on their browsing history. The team has ML expertise but wants to use Google's pre-built ML infrastructure to train and deploy models at scale without managing compute resources. Which Google Cloud service should they use?

Accepted Answer

Vertex AI. Vertex AI is the correct choice because it provides a fully managed, unified ML platform that handles the entire ML workflow—from data preparation and training to deployment and monitoring—without requiring the team to manage underlying compute infrastructure. It integrates with Google Cloud's pre-built ML infrastructure, including distributed training, AutoML, and custom model serving, making it ideal for building and scaling a recommendation engine.

Answer

BigQuery ML

Answer

Cloud AI Platform Notebooks (now Vertex AI Workbench)

Answer

Cloud Dataflow