Ballooned App Base RAM size

[OLG]

HI an amazing Qlik community 👋

For those who have migrated apps to Qlik Cloud - have you come across any major app design offenders that caused a dramatic increase in App Base RAM size (above 5 GB limit)?

We’re seeing an app with a ~0.2 GB file size consuming ~12GB of Base RAM.
Has anyone experienced something similar?

Which visualisations, design choices, data models etc did you find that significantly increased app base RAM ?

Any insights, lessons learned, or best practices would be greatly appreciated 🙏

Qlik Cloud 

[QFabian]

Hi @OLG, in Qlik "consumption" usually refers to the impact on RAM (Memory) and CPU (Processing Power) and disk, With Cloud all of that come togheter as comput unit or GiB in Qlik as measure unit.

To keep your app performing well, you should balance the work done  during the Reload (Script) versus how much calculations your charts does (Expressions).

In script you can check :
Joins, because they can be very dangerous, try mapping or keep
Synthetic Keys and Circular References can slow your apps too.

And in expressions in charts, try to avoid complex, conditionals (if) or nested aggregations, check your set analysis and try to be efficient as much as posible replacing that on the fly calculations for loads in script.

You can check for best practices here in the comunity, for example :

Design

https://community.qlik.com/t5/Design/bg-p/qlik-design-blog

Member Articles

https://community.qlik.com/t5/Member-Articles/tkb-p/qlik-communityarticles

And you can check for Apps behaviour in these Monitoring Apps :

Qlik Cloud Monitoring Apps Workflow Guide :

https://community.qlik.com/t5/Official-Support-Articles/Qlik-Cloud-Monitoring-Apps-Workflow-Guide/ta...

[rwunderlich]

The large base RAM size is going to be related to the data model. Specifically the number of fields and the number of unique values in those fields.

You will want to remove unused fields and optimize the storage of the fields and tables that remain. Here are two blog posts with guides to doing both.

https://qlikviewcookbook.com/2024/11/reducing-qlik-field-memory/

https://qlikviewcookbook.com/2024/05/reducing-qlik-app-resources-with-qsda/

-Rob

[OLG]

Thank you @rwunderlich , @QFabian .Our apps are well‑designed, and the team follows Qlik’s recommended best practices. I suspect that’s the main issue is the size of the data. Is there a recognised standard  ratio between Qlik app file size and base RAM?

Many thanks

Olga

[rwunderlich]

"Is there a recognised standard  ratio between Qlik app file size and base RAM?"

There some rules of thumb bandied about, but those are rules are meaningless in terms of evaluating or optimizing performance of a specific app. 

" I suspect that’s the main issue is the size of the data.".

Precisely. Those articles I pointed you to are intended to help you understand the size of the data and what options you may have for optimizing that size while still retaining the business functionality. Sometimes it's only one or two fields that are taking most of the storage.

-Rob

[rwunderlich]

"Is there a recognised standard  ratio between Qlik app file size and base RAM?"

I'm going to qualify my previous response. The traditional file to RAM ratio bandied about is "10x". As I stated before that just represents a common ratio, not necessarily a good/bad. I've seen this vary quite a bit based on the app data. Your reported 60x (12GB / 0.2GB) is quite unusual in my experience and I'm very curious to know what's causing it. 

-Rob