Cloud Digital Leader Google Cloud Products and Services — Timed 50 Questions

Question 1

A company stores sensitive data in Cloud Storage. They want to restrict access to only users from the company's corporate network (IP range 203.0.113.0/24) and ensure data is encrypted at rest using a customer-managed key. Which TWO configurations are required? (Select 2)

50 minute time limit — choose an answer to begin.

Accepted Answer

Set a bucket policy with an IP address condition. Option C is correct because a Cloud Storage bucket IAM policy can include a condition based on the request's source IP address (for example, using the `request.auth.accessLevels` or an IP-based condition via `resource.name`/`request.remoteIp` style conditions in IAM Conditions), which restricts access to the corporate range 203.0.113.0/24. Option E is correct because customer-managed encryption keys (CMEK) for Cloud Storage are created and managed in Cloud KMS, then referenced by the bucket's default KMS key so data at rest is encrypted with that customer-managed key. Option A is not correct because Cloud Armor protects HTTP(S) load-balanced applications, not direct Cloud Storage bucket access. Option B is not correct because VPC Service Controls provide a service perimeter against data exfiltration but do not by themselves enforce a specific source IP allowlist for bucket access. Option D is not correct because Cloud NAT provides outbound internet access for private instances and has no role in restricting inbound access to Cloud Storage or in encryption at rest.

Answer

Configure Cloud Armor with IP allowlist

Answer

Enable VPC Service Controls

Answer

Configure Cloud NAT

Question 2

A company needs to store petabytes of time-series IoT sensor data and query it with single-digit millisecond latency at millions of reads per second. The data has a simple key-value structure with timestamps. Which Google Cloud database is MOST appropriate?

Accepted Answer

Cloud Bigtable. Cloud Bigtable is designed for exactly this use case — petabyte-scale, low-latency (single-digit ms), high-throughput NoSQL storage for time-series, IoT, and financial data. It scales horizontally by adding nodes. BigQuery is optimised for analytics (seconds-to-minutes latency), Cloud Spanner is a globally distributed relational database for OLTP but not optimised for key-value time-series at this scale, and Firestore is for document data with hierarchical structure.

Answer

Firestore

Answer

BigQuery

Answer

Cloud Spanner

Question 3

A company is deploying a global web application and needs to serve users with low latency, protect against DDoS attacks, and scale automatically. Which two Google Cloud services should they combine? (Choose exactly 2.)

Accepted Answer

Cloud Armor. Cloud Armor (A) is correct because it provides edge security at the Google Cloud load balancer, including DDoS protection via Google's global infrastructure and WAF rules to filter malicious traffic. Cloud Load Balancing (D) is correct because it distributes incoming traffic across backend instances or services globally, supports automatic scaling, and provides the global anycast IP front end that Cloud Armor and Cloud CDN attach to for low-latency delivery. Together they meet the requirements of low latency, DDoS mitigation, and autoscaling. Cloud CDN (B) is not selected because, while it improves latency for cacheable content, it does not itself provide DDoS protection or autoscaling. Cloud NAT (C) is for outbound internet access from private instances, not inbound web serving, and Cloud DNS (E) only resolves domain names and does not provide security or scaling.

Answer

Cloud CDN

Answer

Cloud NAT

Answer

Cloud DNS

Question 4

A company needs to store petabytes of time-series IoT sensor data and query it with single-digit millisecond latency at millions of reads per second. The data has a simple key-value structure with timestamps. Which Google Cloud database is MOST appropriate?

Accepted Answer

Cloud Bigtable. Cloud Bigtable is Google's petabyte-scale, low-latency NoSQL wide-column store designed for time-series and IoT workloads, delivering single-digit millisecond latency at millions of reads/writes per second when the row key is designed for time-series access patterns. Its key-value structure with timestamped rows maps directly to the requirement.

Answer

BigQuery

Answer

Cloud Spanner

Answer

Firestore

Question 5

A company wants to set up a hybrid cloud network connecting their on-premises data center to Google Cloud with high bandwidth and low latency. They also need to provide DDoS protection for their applications. Which THREE services should they consider? (Choose THREE.)

Accepted Answer

Cloud VPN. Cloud VPN provides an encrypted tunnel for connectivity, Cloud Interconnect offers dedicated high-bandwidth, low-latency connections, and Cloud Armor protects against DDoS attacks. Together, these services fulfill the hybrid cloud networking and DDoS protection requirements.

Answer

Cloud CDN

Answer

Cloud VPC

Question 6

A data analyst wants to create interactive dashboards and reports from data stored in BigQuery. They need a free tool that does not require a license. Which TWO tools are appropriate? (Choose TWO.)

Accepted Answer

Google Sheets. Both Google Sheets and Looker Studio are free tools that can connect to BigQuery to create interactive dashboards and reports. Google Sheets offers Connected Sheets for live BigQuery data analysis, while Looker Studio provides a dedicated visualization platform. Dataflow is a data processing service, Looker is a paid BI tool, and BigQuery is a data warehouse.

Answer

Dataflow

Answer

Looker

Answer

BigQuery

Question 7

A company needs to store petabytes of time-series IoT sensor data and query it with single-digit millisecond latency at millions of reads per second. The data has a simple key-value structure with timestamps. Which Google Cloud database is MOST appropriate?

Accepted Answer

Cloud Bigtable. Cloud Bigtable (option B) is correct because it is Google Cloud's fully managed, horizontally scalable NoSQL wide-column store designed for massive time-series and IoT workloads, delivering consistent single-digit millisecond latency at millions of reads/writes per second across petabytes of data with a simple row-key structure. BigQuery (A) is an analytics warehouse optimized for large scans and SQL aggregation, not low-latency point reads at millions of QPS. Cloud Spanner (C) is a strongly consistent relational database that scales well but is not optimized for petabyte-scale key-value time-series throughput at the lowest latency/cost. Firestore (D) is a document database suited to mobile/web app data at smaller scale, not petabytes of high-throughput time-series reads.

Answer

BigQuery

Answer

Cloud Spanner

Answer

Firestore

Question 8

A company uses Cloud SQL for MySQL to store transactional data. They want to run complex analytical queries that join multiple tables and aggregate millions of rows, but the queries are slow and impacting production performance. They need a solution that minimises impact on the transactional database. What is the BEST approach?

Accepted Answer

Export the data to Cloud Storage and load it into BigQuery for analysis. For complex analytical queries involving large aggregations and joins, BigQuery is the optimal solution because it is a fully managed, serverless data warehouse built for this type of workload. Exporting data to Cloud Storage and loading into BigQuery completely offloads analytics from the transactional database, avoiding any performance impact. While Cloud SQL read replicas can offload read traffic, they are not designed for heavy analytical processing and may still suffer from performance issues on complex queries over large datasets.

Answer

Create a Cloud SQL read replica and run analytical queries on the replica

Answer

Upgrade the primary Cloud SQL instance to a higher machine type

Answer

Enable Cloud SQL query caching on the primary instance

Question 9

An engineer needs to deploy a web application that runs on a custom runtime (e.g., Ruby on Rails with system dependencies). The application must automatically scale based on traffic and should not require managing the underlying infrastructure. Which Google Cloud compute service is MOST appropriate?

Accepted Answer

Cloud Run. Cloud Run is the best choice because it is a fully managed serverless platform that can run any container image, including custom runtimes like Ruby on Rails with system dependencies. It automatically scales based on traffic and requires no infrastructure management. App Engine Flexible also supports custom runtimes, but it uses VMs (cannot scale to zero) and is not recommended for new projects; Google advises using Cloud Run instead.

Answer

App Engine Standard environment

Answer

App Engine Flexible environment

Answer

Cloud Functions

Question 10

A company wants to store backup data that will be accessed infrequently (once a quarter) and can tolerate a retrieval time of several hours. The data is critical but regulatory compliance requires it to be retained for 10 years. Which Cloud Storage class is MOST cost-effective?

Accepted Answer

Coldline. Coldline storage is designed for data accessed at most once per quarter and offers lower cost than Nearline/Standard. Archive is cheaper but is intended for data accessed less than once per year, making it unsuitable for quarterly retrieval. The stated retrieval time of several hours is acceptable for both, but access frequency is the determining factor.

Answer

Standard

Answer

Nearline

Answer

Archive

Question 11

A data scientist needs to build a custom machine learning model using TensorFlow on Google Cloud. The training data is stored in Cloud Storage and is hundreds of gigabytes. The data scientist wants to use a managed service that handles hyperparameter tuning and distributed training without managing infrastructure. Which TWO services can be used together to achieve this? (Choose 2)

Accepted Answer

Vertex AI Training. Vertex AI Training (B) is correct because it is the managed service that runs custom TensorFlow training jobs on Google Cloud, supporting distributed training across multiple workers and built-in hyperparameter tuning via HyperparameterTuningJob, all without managing the underlying infrastructure. Vertex AI Workbench (D) is correct because it provides the managed JupyterLab notebook environment where the data scientist develops, tests, and packages the custom TensorFlow training code and then submits it to Vertex AI Training, forming the end-to-end workflow for this scenario. BigQuery ML (A) is not appropriate because it builds models using SQL inside BigQuery on structured data rather than running arbitrary custom TensorFlow code. Cloud Functions (C) and Cloud Run (E) are serverless compute platforms for event-driven or containerized application workloads, not for distributed ML training or hyperparameter tuning.

Answer

BigQuery ML

Answer

Cloud Functions

Answer

Cloud Run

Question 12

A company uses Cloud Functions to process events from Pub/Sub. Which TWO statements about this architecture are correct? (Choose two.)

Accepted Answer

Cloud Functions can acknowledge messages automatically upon successful execution.. Cloud Functions can be triggered by Pub/Sub push subscriptions. When a function is invoked in response to a Pub/Sub message, the service automatically acknowledges the message if the function completes successfully, preventing redelivery. Functions are stateless, short-lived, and scale automatically based on the number of messages.

Answer

Cloud Functions must be associated with a VPC to access Pub/Sub.

Answer

Cloud Functions can run for up to 60 minutes per invocation in all generations.

Answer

Cloud Functions can be used to pull messages from a Pub/Sub subscription synchronously.

Question 13

A company wants to run a batch job that processes large log files stored in Cloud Storage every night. The job typically runs for 2 hours on a single VM with 16 vCPUs and 64 GB of memory. They want to minimize costs. Which compute option is BEST?

Accepted Answer

Compute Engine with preemptible VMs. Preemptible VMs offer up to 80% cost savings over standard VMs and are suitable for fault-tolerant batch jobs that can be restarted if interrupted. Since the job runs nightly for 2 hours and processes logs, it can tolerate preemption with checkpointing. This minimizes cost for the described workload.

Answer

Cloud Run

Answer

Compute Engine with Sole-tenant nodes

Answer

Compute Engine with standard VMs

Question 14

A company uses Google Workspace and wants to analyze employee productivity trends by processing Gmail and Calendar data. They need a solution that integrates natively with Google Workspace and allows custom SQL queries. Which tool should they use?

Accepted Answer

Looker Studio. Looker Studio (option A) is the correct choice because it integrates natively with Google Workspace services like Gmail and Calendar and supports custom SQL queries through its BigQuery connector, enabling productivity trend analysis. It provides a user-friendly visualization layer that can query data directly, fitting the requirement for custom SQL without heavy data engineering. The other options do not fit: Looker (B) is a separate enterprise BI platform that requires a Looker instance and is not natively embedded in Google Workspace, BigQuery (C) is a data warehouse that can run SQL but lacks native Gmail/Calendar integration and visualization, and Dataflow (D) is a stream/batch processing service, not a SQL query or reporting tool.

Answer

Looker

Answer

BigQuery

Answer

Dataflow

Question 15

A company is deploying a stateful web application on Compute Engine that requires persistent block storage for the database. They need high IOPS and the ability to dynamically resize the disk without downtime. Which TWO storage options meet these requirements? (Choose two.)

Accepted Answer

Persistent Disk (SSD). Persistent Disk (SSD) (B) is correct because it is a network-attached block storage option for Compute Engine that delivers high IOPS and can be resized dynamically while attached to a running VM without downtime. Persistent Disk (standard) (D) is also correct because it is likewise a resizable block storage device that supports online growth and can be attached to a running instance, satisfying the persistent block storage and dynamic resize requirements. Local SSD (A) is incorrect because it is ephemeral, physically attached to the host, and cannot be resized or persisted independently of the VM. Cloud Storage (C) is incorrect because it is object storage, not block storage, and cannot serve as a database block device. Filestore (E) is incorrect because it provides a managed NFS file share rather than block storage.

Answer

Local SSD

Answer

Cloud Storage

Answer

Filestore

Question 16

A company needs to store petabytes of time-series IoT sensor data and query it with single-digit millisecond latency at millions of reads per second. The data has a simple key-value structure with timestamps. Which Google Cloud database is MOST appropriate?

Accepted Answer

Cloud Bigtable. Cloud Bigtable is designed for exactly this use case — petabyte-scale, low-latency (single-digit ms), high-throughput NoSQL storage for time-series, IoT, and financial data. It scales horizontally by adding nodes. BigQuery is optimised for analytics (seconds-to-minutes latency), Cloud Spanner is for relational OLTP with strong consistency, and Firestore is for document data with hierarchical structure.

Answer

Firestore

Answer

BigQuery

Answer

Cloud Spanner

Question 17

A company runs a global e-commerce platform that requires a relational database with strong consistency across continents and an SLA of 99.999% availability. Which Google Cloud database service meets these requirements?

Accepted Answer

Cloud Spanner. Cloud Spanner is the only Google Cloud database that provides globally distributed, strongly consistent relational transactions with a 99.999% SLA. Cloud SQL is regional, Firestore is a non-relational document database and does not support SQL relational transactions, and Bigtable does not support SQL and is not strongly consistent.

Answer

Firestore

Answer

Cloud SQL

Answer

Bigtable

Question 18

An engineer needs to create a Pub/Sub subscription that delivers messages exactly once to a subscriber that processes them idempotently. The subscriber runs on Compute Engine and must handle messages at least once but the engineer wants to minimise duplicates. Which subscription type should be configured?

Accepted Answer

Pull subscription with exactly-once delivery enabled. A pull subscription with exactly-once delivery enabled is the Pub/Sub configuration that guarantees messages are delivered exactly once to the subscriber, provided the subscriber acknowledges within the acknowledgment deadline and processes idempotently, which matches the stated requirement to minimise duplicates. Push subscriptions (options B and C) do not support exactly-once delivery, so they can only provide at-least-once semantics regardless of retry policy or extended acknowledgment deadlines. A dead letter topic (option D) only captures messages that repeatedly fail delivery and does not prevent duplicate deliveries, so it does not meet the requirement.

Answer

Push subscription with retry policy

Answer

Push subscription with acknowledgment deadline extended

Answer

Pull subscription with a dead letter topic

Question 19

A company is migrating a legacy monolithic application to Google Cloud. The app has unpredictable traffic patterns and requires custom runtime dependencies (e.g., specific Python libraries with native extensions). The team wants to minimise operational overhead and avoid managing servers. Which compute option is MOST suitable?

Accepted Answer

App Engine Flexible Environment. App Engine Flexible Environment is the most suitable option because it is a managed PaaS that supports custom runtimes via Docker containers, automatically scales to handle unpredictable traffic, and avoids server management. It also supports long-running, stateful behaviors common in legacy monoliths. Cloud Run supports custom containers and scales to zero, but it is stateless and better suited for request-driven, cloud-native services. Cloud Functions is too small/single-purpose, and GKE requires cluster management.

Answer

Google Kubernetes Engine (GKE)

Answer

Cloud Run

Answer

Cloud Functions

Question 20

A company is using Cloud SQL for MySQL and wants to minimise downtime during a schema migration that may take up to 30 minutes. Which THREE steps should they take to achieve near-zero downtime? (Choose 3)

Accepted Answer

Promote the read replica to a standalone instance and redirect traffic. To minimize downtime during a schema migration, create a read replica of the source Cloud SQL instance. Perform the schema changes on the replica (stopping replication if necessary) so the primary remains unaffected. Once the migration is complete and validated, promote the replica to a standalone instance and redirect traffic. This approach avoids downtime on the primary because all changes are applied offline on the replica.

Answer

Use Cloud SQL's Database Migration Service for one-time migration

Answer

Schedule a maintenance window during the migration