Posted On: Jan 7, 2016
You can now configure Amazon Redshift Work Load Management (WLM) settings to move timed-out queries automatically to the next matching queue and restart them. The matching queue has the same Query Group or User Group as the original queue. Please see the WLM Queue Hopping section of our documentation for more detail.
This new feature enables you to route longer-running queries to a separate queue without cancelling them completely. For example, as shown in the screenshot below, you can create two queues—one with a 10-second timeout for interactive queries and another with no timeout for batch queries—and add the same User Group, team1, to both queues. WLM automatically routes any team1 timed-out queries in the interactive queue to the batch queue and restarts them.
