Noos
PT EN
RAN-DBDEA-00277
Ranvier · DBDEA

Exam DBDEA Topics: troubleshooting Difficulty: medium

In the jobs timeline, one stage of a nightly Lakeflow Job takes 40 minutes while its sibling stages finish in about three minutes each. Before changing any configuration, the engineer wants to understand why this particular stage is slow. What is the recommended next step in the investigation?

Options (Reading Mode) Click to filter discussion →
  1. A. Resize the cluster to a larger and more expensive node type right away, since a long stage always means the hardware is undersized
  2. B. Open the stage page and examine the task duration distribution, summary metrics for skew, and spill statistics
  3. C. Assume the notebook code has a bug, because stages only run long when user code contains an error
  4. D. Cancel and resubmit the run, because re-execution is the documented way to clear a long-running stage
Reveal answer

Correct answer: Open the stage page and examine the task duration distribution, summary metrics for skew, and spill statistics

Reference: https://docs.databricks.com/aws/en/optimizations/spark-ui-guide/long-spark-stage

02 — The Myelin Layer · Canonical Discussion

Where the exam tries to trick you — and how those who passed solved it.

0 Passed

contributed to this question

Answer Distractor Matrix Click an option to filter or open its dissection.
✦ Consolidated Synthesis (Wiki) rev. #1 (baseline) · CC BY-SA 4.0
Start synthesis for this question →

No community synthesis consolidated for this question yet. Below is the verifiable official starting point — if you passed the exam, open the option dissection.

Official Answer (B): Open the stage page and examine the task duration distribution, summary metrics for skew, and spill statistics

Ranvier · DBDEA

Contribute to this Question

Your contribution builds your Orbital Sheath

Peer contributions

No public contribution yet. Be the first who passed to explain why the distractors trick candidates — your contribution builds your Orbital Sheath.