In short I am looking for a way to limit QlikView Server memory consumption when too much data has been requested by the user.
We have an inhouse dashboard with multiple tabs and selectors. It can display very detailed data and because of that users abuse it quite often, forgetting to do a selection before diving into detailed analysis. This in turn causes server to run out of memory, we are in the process of amending this by building in a condition into detailed tab - so that it will not display data if no selection has been made (or num of rows to display will be very high).
Problem is that this will take some time to deploy the fix (development cycle takes a while). I was wondering is there a way to limit QlikView Server itself so that it displays 'out of memory' message when users select detailed tab on the dashboard?
Or maybe there is something like 'incremental fetch' functionality?
Because QlikView Server eats all server memory restarting the service takes a while, limiting QlikView service memory would be good as well (is there a way to do that?).
It is healthy behavior if a QlikView server eats up 70% of the memory of a server after some time. There is nothing wrong with that. A common misconception is thinking there is a memory leak or anything wrong.
If users can make pivot tables with a huge amount of dimensions and calculations being used, add a calculation condition to limit this. Though usually these huge things take forever to calculate, so it will run into calculation time-out after a minute anyway.
Are you running a dedicated server? If not, now would be a right time to start doing that .
We run a dedicated server, and including condition is in the works (should be deployed within 1 month).
I was looking for a QlikView software solution, I know it exists in some other BI platforms and was hoping to find something like that in QlikView. I would like qlikview to stop (present user with an error) when Virtual Memory limit has been exceeded. Other option would be to use something like 'incremental fetch' - user would be able to go into 'deep dive' tab, but would be presented with a limited no. of rows (just like queries work in TOAD with default settings).
The problem is that users are careless and crash the server by overloading it.