GCP-PCDE sample questions

Question 1

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?

Accepted Answer

Implement a Spanner autoscaler using a target high-priority CPU utilization of 65%, with maximum nodes set to 15.

Answer

Pre-provision 15 nodes permanently to ensure burst capacity is always available without scaling delay.

Answer

Modify the schema to use bit-reversed sequences for primary keys and scale statically to 8 nodes.

Answer

Migrate the workload to Bigtable, as Spanner cannot efficiently handle write-heavy burst workloads.

Question 2

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?

Accepted Answer

Implement a multi-cluster routing profile that segregates read and write traffic to different clusters within the same instance.

Answer

Deploy a single cluster and vertically scale the nodes to the maximum memory tier.

Answer

Configure the Bigtable cluster with HDD storage to maximize throughput for the 50,000 QPS write load.

Answer

Deploy a single cluster and configure an app profile with single-cluster routing to ensure strong consistency.

Question 3

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?

Accepted Answer

AlloyDB for PostgreSQL utilizing the built-in columnar engine.

Answer

Cloud SQL for PostgreSQL with cross-region read replicas.

Answer

Cloud Spanner using federated queries from BigQuery.

Answer

Cloud Bigtable using Datastream to capture changes.

Question 4

GlobalFin, a multinational financial institution, is re-architecting its core ledger system.

The current system relies on an on-premises Oracle database. The business requires the new cloud-based system to support synchronous replication across continents, guarantee strict external consistency (ACID compliance), and provide a Recovery Point Objective (RPO) of absolutely zero in the event of a total regional failure.

Which Google Cloud database architecture is the ONLY solution that natively meets all these constraints?

```mermaid
graph TD
Client((Client)) --> LB[Global Load Balancer]
LB --> API_US[API Service US]
LB --> API_EU[API Service EU]
API_US --> DB[(Database)]
API_EU --> DB
```

Accepted Answer

Cloud Spanner deployed with a multi-region instance configuration.

Answer

Cloud SQL Enterprise Plus with an Advanced DR replica configured in a secondary continent.

Answer

AlloyDB for PostgreSQL with a primary cluster in the US and a secondary cluster in the EU.

Answer

Cloud Bigtable with a multi-cluster routing profile spanning two continents.

Question 5

A database engineer is designing a disaster recovery strategy for a mission-critical Cloud SQL for MySQL database. The business requires the ability to perform a planned failover (switchover) to a different region with absolutely zero data loss and minimal downtime during regular maintenance drills. Which feature should be implemented?

Accepted Answer

Cloud SQL Enterprise Plus edition utilizing an Advanced DR replica.

Answer

Standard cross-region read replica promotion.

Answer

Regional High Availability (HA) configuration.

Answer

Automated backups stored in a multi-region Cloud Storage bucket.

Question 6

True or False: AlloyDB for PostgreSQL secondary clusters located in different regions use synchronous replication to ensure strict global consistency.

Accepted Answer

False

Answer

True

Question 7

An organization is using a multi-cluster Cloud Bigtable instance to serve a globally distributed application. They want to ensure that reads are always strongly consistent (Read-Your-Writes) for individual users, even if a cluster fails. Which App Profile configuration is required?

Accepted Answer

Single-cluster routing with strict consistency enabled.

Answer

Multi-cluster routing with eventual consistency enabled.

Answer

Any-cluster routing with transactional locks enabled.

Answer

Multi-cluster routing with synchronous replication enforced.

Question 8

A cloud architect is designing the network connectivity for a fleet of 50 Cloud SQL instances across multiple VPCs. The security team mandates that all traffic remain on internal IP addresses. The architect must choose between Private Service Access (PSA) and Private Service Connect (PSC). Which of the following is a critical limitation of PSA that PSC resolves in this scenario?

Accepted Answer

PSA requires reserving large /16 or /24 CIDR blocks which can lead to IP exhaustion, whereas PSC only requires a single IP address per service endpoint.

Answer

PSA traffic is routed over the public internet, whereas PSC keeps traffic entirely on the Google backbone.

Answer

PSA does not support IAM Database Authentication, whereas PSC natively integrates with IAM.

Answer

PSA limits the number of Cloud SQL instances to 10 per VPC, while PSC allows unlimited instances.

Top vendors

Top exams

Free GCP-PCDE sample questions

True or False: AlloyDB for PostgreSQL secondary clusters located in different regions use synchronous replication to ensure strict global consistency.

An organization is using a multi-cluster Cloud Bigtable instance to serve a globally distributed application. They want to ensure that reads are always strongly consistent (Read-Your-Writes) for individual users, even if a cluster fails. Which App Profile configuration is required?

9 more free samples are waiting