PDE Ingesting and Processing the Data Practice Question
A media analytics team runs a Dataflow streaming pipeline that reads click events from Pub/Sub and writes aggregates to BigQuery. During peak hours, the pipeline's BigQuery write throughput plateaus and Dataflow logs show repeated quota-related retries on the streaming insert API. The team wants to keep exactly-once semantics and increase sustained write throughput. What should they change?
⚠ Common exam trap
The trap here is treating a BigQuery API quota plateau as a Dataflow scaling problem and adding workers instead of changing the write path.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Switch the BigQuery sink to the Storage Write API with exactly-once semantics enabled.
The Storage Write API with exactly-once mode is the intended high-throughput sink for Dataflow to BigQuery and removes the legacy streaming insert quota bottleneck while keeping exactly-once delivery. Increasing workers or autoscaling addresses compute, not the API quota, and batching to Cloud Storage trades away latency and complicates semantics. Replacing the sink is the targeted fix.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Enable autoscaling on the Dataflow job and raise the maximum worker count to the regional limit.
Why it's wrong here
Autoscaling adjusts compute based on backlog, but the observed symptom is a BigQuery-side quota, not a CPU or backlog issue. Raising the maximum worker count would not lift the insert quota and could increase contention. The team needs a different write path, not more compute capacity for the same throttled sink.
- ✗
Write the aggregates to Cloud Storage as sharded files and schedule a load job every five minutes.
Why it's wrong here
This introduces five-minute latency and requires an external scheduler, which conflicts with the streaming expectation. It also complicates exactly-once guarantees because the load job must be idempotent against partial files. While it avoids the streaming insert quota, it is a workaround that degrades freshness rather than a direct streaming solution.
- ✗
Increase the number of Dataflow workers to spread the streaming insert load across more threads.
Why it's wrong here
Adding workers does not raise the BigQuery streaming insert quota, which is enforced per project and per table. More workers simply generate more concurrent insert requests that hit the same ceiling, and the retry storm can worsen. Since the bottleneck is the API quota rather than compute, horizontal scaling of the pipeline does not solve the throughput plateau.
- ✓
Switch the BigQuery sink to the Storage Write API with exactly-once semantics enabled.
Why this is correct
The Storage Write API is designed for high-throughput streaming ingestion and, when the exactly-once stream mode is used, provides exactly-once delivery into BigQuery. It avoids the per-row streaming insert quota that causes the plateau while preserving the semantics the team requires. This is the recommended replacement for the legacy streaming insert path in Dataflow.
Go deeper
Related to this question
About these practice questions
One of 747 original PDE practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Google Cloud exam blueprint
This PDE practice question is part of Courseiva's free Google Cloud certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the PDE exam.