Efficient Application Design or Design Best Practise

[HI6]

Hi. I am an Admin for Qlik Sense Enterprise on-premise. Mostly deal with all the application & system backend issues. Not a designer.

Recently we got lots of complains about performance issues and lots of tasks used to abort/fail with no reason.

Then suspected a task, when it loads, CPU utilisation goes to 100% on that Executing node. I monitored that for a while. And found that when that task loads, some other tasks fails.

Now, how I can recommend a good application design best practice guide which gives some guidelines about how to design a efficient application. i.e. SQL world: we talk about normalisations: 1N, 2N, 3N.

Thanks in advance

[Vegar · Accepted solution]

There are several things to consider when designing a Qlik application,  but one general recommendations is to design a star schema model (avoid snowflaking your model). 

Your description indicates issues during reloads. Reload duration is often related to the amount of data processed. You might benefit from looking into the possibility to not process all data every time, but do an incremental load and storing the data in qvd files. Then let your end user application read the QVD files as a source. 

You could also look into the amount of data (which fields and periods do you need) that is loaded in the script. Reducing the time span and the number of fields loaded into your application might also reduce the load on your server. 

You might find some more pointers here: https://help.qlik.com/en-US/sense/May2023/Subsystems/Hub/Content/Sense_Hub/Apps/app-performance.htm and here: https://help.qlik.com/en-US/qlikview/May2023/Subsystems/Client/Content/QV_QlikView/LoadData/best-pra...

Good luck!