The Business Intelligence team reports that when some team members run queries for their dashboards in parallel with others, the query response time is getting significantly slower What can a Snowflak

Question

Exhibits

Options

• AReview the Query Profile in Snowsight to identify bottlenecks and use the AUTO_SUSPEND and AUTO_RESUME features to manage warehouse resources
• BUse the QUERY_HISTORY view in the ACCOUNT_USAGE schema to analyze query execution times and identify resource contention, then consider using multi-cluster warehouses to handle concurrent workloads
• CEnable the Query Acceleration Service for the warehouse and review the WAREHOUSE_METERING_HISTORY to understand usage patterns
• DReview the RESOURCE_MONITOR settings and create separate warehouses for different teams to isolate workloads and prevent resource contention

• A
  3% (1)
• B
  84% (27)
• C
  9% (3)
• D
  3% (1)

Explanation

The symptom - degraded performance specifically when multiple users run queries in parallel - is a classic concurrency/resource contention problem, not a single-query optimization problem. Option B is the most complete and architecturally correct answer: querying ACCOUNT_USAGE.QUERY_HISTORY lets the architect analyze execution times, queuing times, and cluster utilization across all concurrent sessions to confirm resource contention is the cause. The solution - multi-cluster warehouses - directly addresses concurrency by automatically spinning up additional compute clusters when queued queries exceed a threshold, eliminating the bottleneck. Option A (Query Profile + AUTO_SUSPEND/RESUME) is useful for analyzing individual slow queries, not concurrent workload contention. Option C (Query Acceleration Service) helps accelerate specific large analytical scans, not general concurrent dashboard queries. Option D (separate warehouses per team) is a valid isolation strategy but does not solve the concurrency problem for the 2000+ users within the same team, and RESOURCE_MONITOR is a cost-control tool, not a performance diagnostic tool.

No community discussion yet for this question.