Qlik Sense repository service consuming full RAM in the central node

[AkashPohare_HB]

Hi All,

Qlik Sense repository service is consuming full memory in the server and causing node down.

Any idea why would repository service would consume full memory?

Current observation is repository service is gradually consuming full memory in the server. From resource monitor I could see that most memory is consumed while connecting to the dedicated postgres repository database.

Kindly provide any suggestion where should I start checking ?

Thanks.

[Vegar]

I would recommend you to start looking at the Qlik Sense System_Repository log files entries. 

The easiest way to consume these are by using  the Log Monitoring app. Filter on the Service = Repository. Hopefully you will get an indication on what is going on by checking the for a period where you observed the spike. 

Edit: could you also share which version of Qlik Sense and Postgres you are using?

[AkashPohare_HB]

Hello Vegar,

I have checked in the logs, but there is no specific error or information. It has regular operations logged. Any specific thing I should be checking?

Version: Qlik Sense May 2025 Patch 4, standalone postgres repository database.

[Vegar]

@AkashPohare_HB It will be somewhat a guessing game pinpointing your issue. 

You could try to investigate around your logging - do you have exsessinve audit logging that could cause stress to the node? 

You could also try to look at the postgresql activity, if you have some long running tasks. A query like below could be a place to start.

SELECT pid, state, query, now() - query_start AS duration

FROM pg_stat_activity WHERE datname = 'QSR'

ORDER BY duration DESC;

[Lech_Miszkiewicz]

Hi @AkashPohare_HB 

• What does full memory mean for you? 16GB, 64GB or maybe 1.2TB? or other?
• How big is your repository if you do repository backup?
• What version of Qlik and repository (postgres) are we talking about?
• How much disk space do you have on the server

Those are starting point questions I would start with. Back in the day repository could grew for few reasons:

• there was memory leak caused by stale records not being removed from repo - there were sql scripts you could run on postgres repo to clean those up.
• there was a user sync task which would load too many users (often all company with all attributes etc instead of users filtered down to those who should have Qlik access)
• there would be too many users creating too many stories - that was an example of Qlik crashing due to its popularity amongst users years ago.

However without proper investigation it is hard to guess what is causing it in your case.