ADA-C01 Question 407
Single answerDetermine when warehouses should be suspended or resumed based on cost and pricingA company runs an ELT pipeline every 10 minutes using a dedicated Snowflake warehouse. Each run typically executes for 30-40 seconds, and the warehouse then sits idle until the next run. Finance has noticed that warehouse charges are higher than expected. The administrator wants to reduce compute cost without materially delaying the pipeline. Which configuration change is the BEST choice?
- A
Set AUTO_SUSPEND to 60 seconds and keep AUTO_RESUME enabled on the warehouse
- B
Disable AUTO_SUSPEND so the warehouse stays warm between pipeline runs
- C
Increase the warehouse size so each pipeline run finishes faster and total cost decreases
- D
Keep AUTO_SUSPEND at 10 minutes to avoid paying for warehouse restarts
Show answer and explanation
Correct answer: A
Explanation
For warehouses with intermittent activity, Snowflake best practice is to use AUTO_SUSPEND and AUTO_RESUME to minimize charges during idle periods while preserving operational convenience. In this scenario, the workload is bursty and predictable: short processing windows followed by long idle gaps. The best action is to suspend the warehouse shortly after work completes and allow automatic resume at the next execution. Administrators should base suspend/resume decisions on the relationship between active processing time, idle time, and acceptable startup latency. Snowflake documentation and cost-management guidance consistently recommend enabling AUTO_SUSPEND and AUTO_RESUME for warehouses that are not continuously busy, especially when idle periods are longer than the small delay associated with resuming a suspended warehouse.
- A. Correct.
Correct. Because the workload runs briefly every 10 minutes, the warehouse is idle for most of the interval. Setting AUTO_SUSPEND to 60 seconds allows Snowflake to stop billing for idle time soon after each run completes, while AUTO_RESUME lets the next scheduled run start the warehouse automatically. This is the standard cost-optimization pattern for intermittent workloads. Snowflake charges warehouses while they are running, so reducing idle runtime is the main lever here.
- B. Incorrect.
Incorrect. Leaving the warehouse running between short, infrequent jobs causes unnecessary charges during the 9+ minutes of idle time between executions. Some administrators choose this to avoid startup latency, but for a workload that only runs every 10 minutes, the added idle cost typically outweighs the small resume delay.
- C. Incorrect.
Incorrect. A larger warehouse may reduce query runtime, but it also consumes credits at a higher rate while running. For a short pipeline with long idle gaps, the primary cost issue is idle time, not insufficient compute capacity. Upsizing the warehouse does not address the wasted credits from leaving it active between runs.
- D. Incorrect.
Incorrect. A 10-minute AUTO_SUSPEND setting means the warehouse would likely remain running for the entire interval between jobs, effectively billing continuously even though work only occurs for 30-40 seconds. The idea that longer suspend times avoid restart cost is a common misconception; the bigger cost concern here is paying for idle warehouse uptime.