Google Professional Data Engineer exam dumps

Google Professional Data Engineer practice question 229 of 279

Professional Data Engineer. Professional level, Google Cloud. Free question with the correct answer and a full explanation.

Google Professional Data Engineer Question 229

Single answerGoogle Cloud Platform

You are designing a data processing pipeline for your company's batch and ad-hoc analytics workloads using Google Cloud Dataproc. The batch jobs are scheduled to run every night and can take several hours, while ad-hoc jobs are submitted sporadically during working hours. Cost efficiency and operational simplicity are top priorities. How should you configure your Dataproc clusters?

  1. A

    Use a single persistent Dataproc cluster for both batch and ad-hoc workloads to avoid cluster provisioning delays.

  2. B

    Use a persistent Dataproc cluster for batch workloads and create ephemeral (job-based) clusters for ad-hoc workloads.

  3. C

    Use ephemeral (job-based) clusters for both batch and ad-hoc workloads.

  4. D

    Use two separate persistent Dataproc clusters: one for batch workloads and another for ad-hoc workloads.

Show answer and explanation

Correct answer: B

Explanation

The best solution balances cost efficiency and operational simplicity. A persistent cluster is ideal for predictable, scheduled batch jobs to avoid provisioning delays, while ephemeral clusters for ad-hoc workloads ensure resources are only provisioned when needed, reducing idle costs.

  • A. Incorrect.

    Using a single persistent Dataproc cluster for both workloads simplifies management but may lead to underutilization during idle periods, increasing costs. It also risks resource contention between batch and ad-hoc jobs.

  • B. Correct.

    Using a persistent cluster for batch workloads ensures availability for scheduled jobs, while ephemeral clusters for ad-hoc jobs minimize idle costs and provide flexibility by provisioning only when needed.

  • C. Incorrect.

    Using ephemeral clusters for both workloads eliminates idle costs but could introduce delays for batch jobs due to cluster provisioning time, which is less suitable for regular, time-sensitive schedules.

  • D. Incorrect.

    Using two separate persistent clusters eliminates resource contention but significantly increases costs due to the always-on nature of both clusters, even when idle.

Timed practice exam

Take a Google Professional Data Engineer practice test under exam conditions

60 questions in 120 minutes, drawn from this bank, with a score report and a per-question review when you finish.

Start timed exam