topic Re: QIX performance log NetRam and PeakRam in QlikView

QIX performance log NetRam and PeakRam

roeldh — Wed, 25 Nov 2020 16:16:04 GMT

Hey everyone,

I was trying to use the QIX performance log to see if I could get the exact ram usage at any given time.

I was going to use this to anaylese the ram usage of some dashboards.

But I found that the NetRam went negative and started questioning what the definition of the NetRam really was within QlikView.

I also thought the PeakRam would be the total used Ram by the system at a given time but since there multiple values at 1 given timestamp with highly changing values this doesn't seem to be true either.

So can someone pls give me the correct definitions of the PeakRam and NetRam that QlikView use?

When is the NetRam and PeakRam messured at the start of a thread during or end of it?

Thank You, for taking time to look at this.

Re: QIX performance log NetRam and PeakRam

rubenmarin — Tue, 23 Apr 2019 17:24:13 GMT

Hi, like to help definitions: https://help.qlik.com/es-ES/qlikview/November2018/Subsystems/Server/Content/QV_Server/QlikView-Server/QVSRM_Logging.htm#anchor-10

It doesn't explains the negative NetRam

Re: QIX performance log NetRam and PeakRam

Daniele_Purrone — Wed, 24 Apr 2019 14:21:11 GMT

Hi!
This was reported in Qlik Sense, and it's currently being investigated as a bug there: https://support.qlik.com/articles/000072892

The Engine component is common between the two products, so we can expect the investigation to be beneficial for QlikView as well.

Re: QIX performance log NetRam and PeakRam

roeldh — Thu, 25 Apr 2019 06:33:52 GMT

Oke thank you this does explain the net RAM going negative.

But a part of my question does still remain. That would be what is the PeakRAM within Qlik?

As far as I remeber the PeakRAM is the total RAM beeing used by the full system but having multiple values at 1 time would make that impossible.

Re: QIX performance log NetRam and PeakRam

Brett_Bleess — Tue, 18 Jun 2019 16:05:45 GMT

I ran some tests myself just to see as I have not messed with this feature since the presentation I did on the new settings a number of months back, but from what I can see, the Peak RAM is the maximum RAM used to handle that specific request, so each line of the log will have a different value except for instance the custom protocol requests seem to be pretty much the same value every request, which does make sense. Hopefully this helps a bit more, but the key is the Method column, that is what things are really tied to here, so hopefully this does make a bit more sense, so the Peak RAM will be the peak memory used to process that Method call I believe.

Regards,
Brett