QuestionQ49

Performance, Cost, and Resource Optimization

An Architect must improve the performance of reports that retrieve data from multiple Snowflake tables, join it, and then aggregate it. Users access the reports through several dashboards. Performance problems occur on Monday mornings from 9:00am–11:00am when many users review the sales reports.

The group size has grown from 4 to 8 users. Dashboard refresh wait times have increased substantially. This workload is currently served by a virtual warehouse with these parameters:

  • AUTO-RESUME = TRUE
  • AUTO_SUSPEND = 60
  • SIZE = Medium

What is the MOST cost-effective way to increase report availability?

Explanation

Multi-cluster warehouses in auto-scale mode are designed to manage fluctuating user and query concurrency. With a minimum of one Medium cluster and a maximum of four, Snowflake can start additional clusters when load causes queries to queue and shut them down as load falls, improving dashboard-refresh availability without continuously paying for all clusters. Snowflake recommends auto-scale mode unless there is a specific need for maximized mode.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!