QuickSight makes separate query to the datasource for every widget in the dashboard

We have noticed that QuickSight currently makes separate queries to the data source for each widget in our analysis or dashboard. However, it would be more efficient if QuickSight could make a single query to the data source to fetch all the required data within specific filters, and then perform the remaining operations internally.
Please let us know if this can be optimised further.

Hi @sohail - Is it possible to use SPICE for your use case? In that case whole data set will load to SPICE and all your visualization and analysis query will hit to SPICE instead of databases. However you have to explore an option to refresh the SPICE depending upon database tables refresh.

Hi @David_Wong - kindly provide your feedback on this?

Regards - Sanjeeb

1 Like

Hi @sohail, @Sanjeeb2022,
I would also suggest using SPICE instead of direct query.

1 Like

We are planning to move to SPICE. Would it resolve this issue?
Also please help me understand why is it making separate query calls, i mean what is the mental model behind this.

Hi @sohail - SPICE will surely improve your performance and all the queries will hit the SPIC data set and not go to database. Now why the queries are running in parallel, this a design discussion and something AWS followed in the backend. If you are looking more details, please submit a ticket to AWS Customer support team.

Regards - Sanjeeb

@sohail ,
Each visual in QuickSight can have different configurations ( columns , filters ) . When a query is sent to the database, it will only include the required columns with corresponding filters and bring the results back.

As @Sanjeeb2022 and @David_Wong said, you can use SPICE . It is designed for query performance, concurrent usage which means your backend database is not impacted by many users accessing and using the dashboard interactively.
Note : SPICE being a snapshot of your data, you have to schedule refreshes to update data.

I am marking this response as solution.

Kind regards,
Koushik

1 Like