GCP-PCDE Google Cloud Professional Cloud Database Engineer Practice Questions
Prepare for GCP-PCDE with more than an answer.
Unlock the full exam and previous versions
- v1Google Cloud Professional Cloud Database Engineer 207 questions Current
- PCDBELegacy Professional Cloud Database Engineer 258 questions Locked
- Exam fee
- $200 USD
- Level
- Professional
- Valid for
- 2 years
Domains covered on the exam 4
- Design Innovative, Scalable, and Highly Available Cloud Database Solutions32%
- Manage a Solution That Can Span Multiple Database Technologies25%
- Migrate Data Solutions23%
- Deploy Scalable and Highly Available Databases in Google Cloud20%
- 1
A machine learning engineering team is building a Retrieval-Augmented Generation (RAG) application. They need to store 500 million high-dimensional vector embeddings and perform similarity searches. The application requires extreme performance, specifically demanding that vector searches complete up to 10x faster than standard PostgreSQL pgvector indexes. Which database technology should you recommend?
Show answer details
Correct answer: C
AlloyDB integrates Google's proprietary ScaNN (Scalable Nearest Neighbors) algorithm into its pgvector extension. This specific integration allows AlloyDB to perform vector similarity searches up to 10x faster than standard open-source PostgreSQL pgvector indexes, making it ideal for high-performance RAG workloads.
- 2
A manufacturing enterprise is migrating to Google Cloud. Their core ERP system runs on an Oracle Real Application Clusters (RAC) database. The database uses proprietary Oracle features, requires sub-millisecond latency to local storage, and the vendor refuses to certify the application on any database other than Oracle. Which Google Cloud solution is required for this workload?
Show answer details
Correct answer: C
Bare Metal Solution (BMS) provides purpose-built, dedicated hardware collocated in Google Cloud data centers. It is specifically designed for lift-and-shift migrations of specialized workloads like Oracle RAC that require hardware-level access, specific hypervisors, and strict vendor certification support.
- 3
A development team is building a new mobile application that allows users to collaborate on shared documents in real-time. The application must support offline mode, automatically synchronizing local changes with the cloud database when network connectivity is restored. The schema will be highly flexible and hierarchical. Which database service is most appropriate?
Show answer details
Correct answer: C
Firestore is a NoSQL document database specifically designed for mobile and web applications. It natively supports a flexible hierarchical schema, real-time listeners for collaboration, and built-in offline synchronization via its mobile SDKs.
- 4
True or False: When architecting a high-throughput caching layer to reduce read load on a primary Cloud SQL database, Cloud Bigtable is the recommended solution because of its sub-10ms read latency.
Show answer details
Correct answer: B
False. While Bigtable offers sub-10ms latency, it is a persistent NoSQL wide-column store, not an in-memory cache. Memorystore (Redis or Memcached) is the recommended GCP service for caching, offering sub-millisecond latency specifically designed to alleviate read load on primary databases.
- 5
A developer needs to connect a microservice running on Google Kubernetes Engine (GKE) to a Cloud SQL for PostgreSQL database. Following the principle of least privilege, what is the most secure method to grant this access?
Show answer details
Correct answer: C
Workload Identity is the recommended way to securely authenticate GKE workloads. By binding a specific KSA to a dedicated GSA with only the
Cloud SQL Clientrole, and utilizing IAM database authentication (passwordless), you strictly enforce the principle of least privilege. - 6
An e-commerce platform relies on Cloud Spanner to handle extreme write-heavy burst traffic during flash sales. The current configuration uses 5 nodes. During bursts, CPU utilization hits 95% and write latencies spike unacceptably. Cost optimization is strictly required during non-burst periods. Which approach provides the optimal balance of performance and cost?
Show answer details
Correct answer: A
Spanner autoscaling dynamically adjusts compute capacity based on CPU utilization. Setting a target of 65% for high-priority CPU ensures sufficient headroom to absorb sudden spikes (preventing the 95% saturation) while automatically scaling down during quiet periods to save costs.
- 7
A data analytics team is sizing a new Cloud Bigtable cluster for a real-time IoT ingestion pipeline. They have strict Service Level Agreements (SLAs) dictating that 99th percentile read latency must remain under 10ms. During peak ingestion, write throughput will reach 50,000 QPS. Which architectural decision best ensures these SLAs are met while optimizing cluster sizing?
Show answer details
Correct answer: B
Workload isolation using app profiles is a best practice in Bigtable. By directing heavy write traffic to one cluster and latency-sensitive read traffic to another cluster within the same instance, you prevent write-induced CPU saturation from impacting the 10ms read latency SLA.
- 8
A retail organization is migrating its inventory management system to Google Cloud. The workload is a Hybrid Transactional/Analytical Processing (HTAP) scenario, requiring high-throughput ACID transactions alongside complex, real-time analytical queries over the same data. Which database service and specific feature should be utilized to satisfy these requirements without impacting transactional performance?
Show answer details
Correct answer: C
AlloyDB is designed specifically for HTAP workloads. Its columnar engine automatically keeps frequently queried data in a columnar format in memory, drastically accelerating analytical queries without degrading the performance of concurrent transactional inserts and updates.
