Question 560
For this Snowflake situation — A warehouse was increased two sizes but end-to-end data delivery barely improved because upstream file arrival dominates the delay. — the team adopts: Measure each pipeline stage before changing compute again and optimize the true critical path. Which ongoing practice BEST limits the same problem recurring?
Explanation
Correct Answer
The relevant ongoing safeguard is: Keep pipeline-level service metrics in addition to warehouse query timing. It controls the recurring exposure associated with this mechanism: End-to-end latency can be limited by waiting for source data rather than SQL execution speed. The original action, measure each pipeline stage before changing compute again and optimize the true critical path., still needs operational checks when the environment or workload changes.
Incorrect Answers
The suggested review tracks a different technical risk: Scaling out adds clusters for parallel query traffic instead of only making one cluster bigger. Here the sustained control must address: End-to-end latency can be limited by waiting for source data rather than SQL execution speed. Choose “Keep pipeline-level service metrics in addition to warehouse query timing”.
That safeguard is relevant to: QUALIFY filters rows based on window-function results after analytic evaluation. The current failure instead follows from: End-to-end latency can be limited by waiting for source data rather than SQL execution speed. Its prevention measure is “Keep pipeline-level service metrics in addition to warehouse query timing”.
The suggested review tracks a different technical risk: Query Acceleration Service can offload eligible scan-heavy work to shared compute, subject to feature criteria. Here the sustained control must address: End-to-end latency can be limited by waiting for source data rather than SQL execution speed. Choose “Keep pipeline-level service metrics in addition to warehouse query timing”.