SnowPro Associate: Platform Question 11
Single answer● Elastic computeA data engineering team runs large ELT transformations every hour, while BI analysts run dashboards continuously throughout the day. Both workloads currently use the same Snowflake virtual warehouse. During business hours, analysts report slow dashboard performance whenever the hourly ELT jobs start. The company wants to improve concurrency for dashboards without changing the underlying table data and while preserving Snowflake's separation of storage and compute. Which action is the best solution?
- A
Create a separate virtual warehouse for BI queries and keep the ELT jobs on their own warehouse
- B
Increase the data retention period on the database so dashboard queries can use historical versions without contention
- C
Convert the warehouse to a reader account so compute can be isolated for each workload
- D
Enable Time Travel on the database so ELT and BI queries no longer compete for compute resources
Show answer and explanation
Correct answer: A
Explanation
Snowflake provides elastic compute through virtual warehouses, which are independent compute clusters that can access the same centralized storage layer. Because compute and storage are separated, the recommended way to prevent one workload from affecting another is to isolate them on different warehouses. In this scenario, placing BI and ELT on separate warehouses improves performance predictability and concurrency without duplicating data. This aligns with Snowflake best practices for workload isolation, where different workloads such as ETL, reporting, and data science commonly use dedicated warehouses sized and scaled to their needs. Features like Time Travel and data retention relate to historical data access and recovery, not compute isolation.
- A. Correct.
Correct. Snowflake's elastic compute model allows independent virtual warehouses to access the same stored data without copying it. Assigning BI and ELT to separate warehouses isolates compute resources, so heavy ELT processing will not consume the same warehouse resources used by dashboard queries. This is a core best practice for workload isolation and concurrency management in Snowflake.
- B. Incorrect.
Incorrect. Data retention controls how long historical table versions are kept for Time Travel and Fail-safe-related recovery use cases. It does not improve query concurrency or isolate compute resources between dashboards and ELT jobs. Someone might choose this option by confusing data versioning features with workload performance features.
- C. Incorrect.
Incorrect. Reader accounts are used to share data with consumers who do not have their own Snowflake account. They are not a mechanism for isolating internal workloads such as BI and ELT running in the same account. This option reflects a misunderstanding of data sharing versus warehouse design.
- D. Incorrect.
Incorrect. Time Travel allows access to historical data states, but it does not separate compute consumption between concurrent workloads. BI and ELT queries would still compete if they run on the same warehouse. This distractor targets the common misconception that storage features can resolve compute contention.