Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Database Engineer practice bank has an additional 769 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
Your team is migrating a production Amazon RDS for PostgreSQL database to Cloud SQL for PostgreSQL. You are evaluating current workload metrics to establish a performance baseline and size the target environment:
Which approach should you take to establish an accurate performance baseline and size the target Cloud SQL instance?
This approach relies on mean utilization values over a 30-day period and attempts a 1:1 hardware mapping (sometimes called a no-rightsizing strategy) from the source Amazon RDS instance shape to a Cloud SQL machine type, while leaving storage sizing unchanged.
The workload maintained completely flat, uniform utilization 24/7 with minimal disk I/O and network demands, and you were performing a rapid, non-rightsized initial TCO estimate rather than production performance sizing.
Provisioning storage strictly based on used data volume minimizes disk capacity costs, while assuming burst IOPS mechanisms (such as AWS gp2 or gp3 credit buckets) will handle temporary spikes in read and write demand.
You were sizing a low-traffic development or archival environment with minimal active queries, where minimizing storage cost is the top priority and high I/O latency is acceptable.
Establishing a migration performance baseline requires capturing resource consumption across distinct workload profiles—such as peak daytime OLTP and nightly batch processing—and translating those measurements into Google Cloud infrastructure limits. In Cloud SQL for PostgreSQL, storage performance (IOPS and disk throughput) as well as maximum network egress bandwidth scale directly with both the number of vCPUs and the provisioned disk size.
Unlike relying on 1:1 hardware matching or 30-day averages, evaluating peak-load profiles alongside joint vCPU-and-disk scaling guarantees that the target Cloud SQL instance sustains source SLAs without performance degradation or migration timeouts.
Adding multiple network interfaces (NICs) or additional IP addresses to a virtual machine or managed database host connects it to multiple VPC networks for routing segmentation or security isolation, rather than expanding its physical or virtual bandwidth capacity.
Your primary goal was isolating database management, backup, and client traffic across separate VPC networks using dedicated private IP interfaces, while already having a machine type sized for peak egress.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated