CERTIFIED-DATA-ANALYST-ASSOCIATE Question #35: Real Exam Question with Answer & Explanation

Question

A data analyst has set up a SQL query to run every four hours on a SQL endpoint, but the SQL endpoint is taking too long to start up with each run. Which of the following change can the data analyst make to reduce the start-up time for the endpoint while managing costs?

Options

• AReduce the SQL endpoint cluster size
• BIncrease the SQL endpoint cluster size
• CTurn off the Auto stop feature
• DIncrease the minimum scaling value
• EUse a Serverless SQL endpoint

Explanation

Serverless SQL endpoints eliminate cold-start latency because Databricks pre-provisions and manages the underlying compute infrastructure on your behalf - when your query fires every four hours, the endpoint is ready in seconds instead of waiting minutes for cloud VMs to spin up. Serverless also manages costs automatically since you only pay for active query time, satisfying both requirements in the question.

Why the distractors are wrong:

• A (Reduce cluster size): Smaller clusters may marginally reduce startup overhead but don't solve the fundamental cold-start problem and sacrifice query performance.
• B (Increase cluster size): More nodes actually means longer startup time, the opposite of what's needed.
• C (Turn off Auto stop): This eliminates restarts by keeping the endpoint running 24/7, but you'd pay for three to four hours of idle compute between each run - poor cost management.
• D (Increase minimum scaling): This affects how many parallel clusters are available for concurrency/scaling, not how fast a single endpoint cold-starts.

Memory tip: Think "Serverless = no server wait" - Databricks holds warm infrastructure in reserve so your workload never waits for a server to boot. The other options either slow startup, waste money on idle time, or address scaling rather than startup.

No community discussion yet for this question.