A virtual warehouse initially experienced poor performance due to queries from multiple concurrent processes queuing. The problem resolved itself over time.

What action can be taken to prevent this from happening again?