Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
Hi everyone
I've recently developed a mashup and started testing its performance. I've noticed that there is a major difference between useres.
The users with high privileges (Developers, Admins) are getting a much better performance than the "normal" users, that suffers much slower app. (not only the first load of the mashup. also moving between pages in the mashup).
I started investigating the issue and I noticed that every time a user starts a session (Mashup or Normal Connection through the Hub), there are 8-15 rows in the System Engine Logs (Engine\Trace\HostName_System_Engine ) that says "QrsSecurityManagment: Retrived Privileges in X ms From the QRS" - The "X" represent a number that's different between users.
The interesting issue here is that when I'm (or other user) accessing the app through the Hub, threre are 10 - 15 " QrsSecurityManagment: Retrived Privileges in X ms From the QRS" rows in the log for the specific session, but when we're accessing the app through the mashup there are around 250 (!!) rows.
For me (rootAdmin) and the qlik developers (wide access), the "retrieving from the QRS" operation takes 8 - 20ms.
for the "Normal" users (clients) this operation takes 60-80 ms.
im guessing the 250 rows, multiple the difference in the length of the operation of retrieving privileges is what's causing the performance issue for the normal users.
anyone knows the "QrsSecurityManagment: Retrived Privileges in X ms From the QRS" operation and its dependencies?
what could we do that made this mashup cause all of those calls?
important to know - that's not our first mashup on our qs site, but it's the first time we experiencing this issue.
thanks
Hi Everyone,
Thanks for your reply.
we are killing all of the hypercubes as explained in the video.
After consulting qlik's Support we understood that in this specific mashup we had a big amount of objects (around 100), and that was the trigger for all of the QRS calls.
Hi @kobidaniel112 ,
Have you tried closing/killing the objects when it's done it's part of the example video i am sharing please take a look.
Hi @ajaykakkar93 ,
Thank you for the video, very useful.
I have a question. I have an app with 4 pages. The app is heavy because it is using a large amount of data. I don't close my hypercubes at the moment. So every time I am selecting new filters, it can take some time to refresh all the charts, but once I have them, navigating through the different pages is very quick.
Should I close the hypercubes each time I switch page, will that be longer to navigate between pages ? I mean is there a priority between having a short page loading time for each page or a long loading time at first (and every new selection) but a smooth navigation ?
It is my first mashup so I am not quite sure what is the best thing to do.
Thank you,
Leo
Hi @LCauch ,
I think closing the Hypercube will be the best approach in terms of long usage & not the best but good performance for the users & if this approach is followed it will be applied for future versions of qlik as well without any changes to the existing mashup.
Hi Everyone,
Thanks for your reply.
we are killing all of the hypercubes as explained in the video.
After consulting qlik's Support we understood that in this specific mashup we had a big amount of objects (around 100), and that was the trigger for all of the QRS calls.
Hi @kobidaniel112 - Did it help to improve the performance then ?
Hi @ajaykakkar93 - Thanks fr sharing video, cleared many doubts.