Btw, I checked, and it's about 80GB.
And in fact, it could be possible that with compression and encoding, the core data in the memory of Qlik Sense is just 30GB.
I accept your answer as correct because it just makes sense considering the RAM issue.
But I have 256 Core, 256GB of RAM, 20GB of network speed and in some Sheet, I still wait for a few seconds to rebuild the chart, after the selection.
I'm using EC2 a m4.16xlarge, I'm still figuring out what is the best solution to speed up Qlik Sense Server.
I have a demo with my best client in a couple of hours and I could boost the server.
What I could increase or optimize to get more speed. I'm on AWS so that I can change configuration with a restart (a few minutes).
From External machine or internal machine is the same.
I'm connecting with a 100Mbs fiber to the server.
I can see the loading icon spinning for a few seconds (5-20secs).
I noticed that is happening the first time, I enter a sheet. If I go back/forward and step back the is loading immediately (cache effect). But if I close and re-open the browser, I can still see the effect of the cache: a page loaded before pops up quickly. So I assume that the cache is on server side.
On Saturday maintenance windows:
I made QSE swallow 500GB of raw data from RedShift.
QSE is using all the available RAM now.
It's impressive the efficiency of Qlik to squeeze data in RAM.
I noticed that with large dataset the waiting for Data loading is increasing. I clicked in Data Manager and more than 10 minutes to just get the tables list. I will study to optimize this, at least, I will open a new thread.