How does quicksight direct query effect redshift load based on number of users?


I have a dashboard which has 500 gb data. I have kept it in direct query mode. There are 100 users who are consuming the dashboard. How does this impact my redshift load? I have directly connected redshift tables to the quicksight dashboard.

Does the query scan based on each user for every filter that apply by every user?

Can you help me with this please?

Hi @salagalakiran123 - It depends upon the database tables configurations and type of sql query. Depending upon the number of users ( concurrent users) and type of sqls the redshift load will be affected. You can monitor the performance of the redshift load by the QuickSight user as well. Please see the below documentation link - Working with performance data in the Amazon Redshift console - Amazon Redshift

Possibly you need to explore WLM and set up a dedicate queue for all queries fired by QuickSight so that you can balance the redshift cluster load.

Regards - Sanjeeb


Thank you Sanjeeb for the information. I understand we need to manage the WLM and depends on the database tables configuration. But, my question is, does quicksight raise queries each time by every user when they access the direct query dashboard? Is it yes or no? Could you confirm this please. When I tested in real time, even if I change the filters in the dashboard, quicksight user triggers the query to the database. I see that around 1.5PB data scan is happening due to which my redshift has reached to 100% CPU

Hi @salagalakiran123 - Yes, when you are selecting “Direct Query” method in QuickSight, essentially user is requesting QuickSight to connect to the target and execute the query and provide the update results. So every time it will hit the database.

One of the option you can explored SPICE. If the data is refreshed in batch mode, you can bring the cut off ( filter version of the data) to SPICE and this way you can avoid hitting the data to database.

Regards - Sanjeeb

1 Like