A data team is using a Snowflake virtual warehouse for both ETL processes and ad-hoc reporting. During peak ETL hours, reporting queries are significantly slowed down. What is the BEST strategy to mit

Question

Options

• AIncrease the size of the existing warehouse to X-Large.
• BCreate two separate virtual warehouses: one for ETL and one for reporting.
• CSchedule all reporting queries to run outside of ETL hours.
• DReduce the concurrency level of the ETL processes.
• EEnable resource monitors to limit the credit usage of reporting queries.

• A
  11% (3)
• B
  63% (17)
• C
  7% (2)
• D
  19% (5)

Explanation

Creating separate virtual warehouses for ETL and reporting (B) is the best strategy because Snowflake's compute is isolated per warehouse - dedicated warehouses eliminate resource contention entirely without wasting credits on always-on over-provisioning. A (upsizing to X-Large) throws money at the symptom rather than the cause; both workloads still compete on the same compute, and you pay for peak capacity 24/7. C (scheduling around ETL) solves the conflict by removing flexibility - it forces business users to wait and doesn't scale as data grows. D (reducing ETL concurrency) trades one problem for another: ETL slows down to let reporting breathe, but you've just shifted the bottleneck. E (resource monitors) controls credit spend, not query performance - capping credits won't prevent reporting queries from being queued behind ETL.

Memory tip: Think "separate concerns, separate warehouses." In Snowflake, compute is cheap to spin up and scales independently, so the idiomatic solution to workload isolation is always splitting warehouses, not throttling or scheduling around contention.

No community discussion yet for this question.