Embedded Quicksight Dashboard loading too slow

Q&A
, ,
1

I have about 10 dashboards in QS and embedded using generated code in Lambda function to the frontend Vercel Webapp created using React.js, client access authentication via Cognito.
There is a number of calculated fields in the dashboard analysis backend, the issue is it takes about 5 mins to load the visualization in the webapp, same with the dashboard and analyses in Quicksight. There are 7 filters and 8 charts per sheet.

Is there a way to speed up the loading time of the Vizs?
Thanks in advance

2

Hi @Tes,

Are you using SPICE or direct query?

3

Hello @David_Wong

SPICE connection to Athena.

4

Do you have a large number of calculated fields? Have you tried moving them from the analysis to the dataset? I think that should help.

5

Not sure how many is considered large regarding the calculated fields.
Varies between 7 and 10, per analyses page
Each analysis and corresponding dashboard, created separately but using the same dataset.

What do you mean move them to dataset?

6

When you’re creating/editing a dataset, you have the option to create calculated fields. Can you create your calculated fields that way instead of doing it inside your analysis?

image
image490Ă—263 12.7 KB

Quick Sight then will perform the calculations when it refreshes the dataset as opposed to doing it when loading your visuals.

You can also check out this blog:

image
image796Ă—389 41 KB

2 Likes
7

Thank you @David_Wong . This is a very performance tuning approach to move the calculated field to the data preparation stage rather creating it at analysis level.

Hi @Tes - Is it possible to access the dashboard and analysis directly from Quick Sight ( URL generated by the app) to see the performance.

Regards - Sanjeeb

1 Like
8

Hi @Tes, did this solution work for you? I am marking this reply as, “Solution,” but let us know if this is not resolved. Thanks for posting your questions on the Quick Sight Community!

9

Thank you very much @David_Wong , this has really been helpful.

Does that mean I cant move my other calculations involving LACs, like SumOver with PreAGG and PreFilter, to the data prep/ edit dataset stage?; because the calculation results I am getting are now different.

10

@Tes
If your LACs are not giving you the expected results when created at the dataset level, I would suggest opening a separate topic to describe the differences that you see when the LACs are created at the dataset level versus at the analysis level.

1 Like
11

Did this solution work for you? I am marking this reply as, “Solution,” but let us know if this is not resolved. Thanks for posting your questions on the Quick Sight Community!