Further information

See graph below. Around 9:15 on 16 Jan, the Problems tabs on both hosts have been closed. Only the dashboards were then running on both hosts. It can be observed a tiny release of memory after closing the Problems tabs but most of the memory remain used and the memory utilization continues to increase with only the dashboards running. At 8:00 on 17 Jan, both Firefox and Chrome browsers have been closed, the memory has been released.

Hello Olivier, what OS are those browsers running on?

Hi Edgars, both are running on CentOS 7.7

Can you add an additional item, proc.mem["processname"] to the hosts? For both, chrome and firefox?

Edgars, this time I only let the dashboard running on both hosts, no additional Problems tabs.

We see the same behavior at memory utilization

And below the proc.mem graph including chrome process (host 1) and firefox + Web Content processes (host 2)

@zux : any news concerning this issue ? Do you need more inputs ?

@zux : what is the status of this issue ?

@zux : can you please provide a status about this issue ? The status is "Need Info" for weeks now but the requested information has been provided. I would like to know if you're going to look deeper in this problem or not. Thanks.

Hello Olivier, sorry for the long wait. We will try to look into this deeper, but we're currently unable to reproduce the issue.

Hello,

I can't reproduce at home, but at work, I encounter this kind of problem with all my setups ( 4.4 and 5.0.1 ) and since a few months

Browser on Win 10 with different setups ( plugins / installed soft )

I noticed it when I have left a chrome tab on the Monitoring/Dashboard/Global View over night.

This chrome tabs end with "not enough memory" when I get back to it.

So I created a new Dashboard with only one widget (on the 5.0.1). Widget Problems / Refresh 10 sec (to show the pb faster) / Show History / show lines 100

I went to the "more tools/task manager" within chrome to get the memory consumption of the dashboard tab

I also "inspect" the page and went to the "Network tab".

I could notice that for every refresh of the widget, 700Kb of data was fetched with  zabbix.php?sid=xxxxxxxxxxxx&action=widget.problems.view

I could also see a bump in the memory consumption of the tab until the tab crashes over night.

I could see this happen with different version of chrome ( version are not older than 3 months ) and different versions of the fronted (4.4+ and 5.0+)

Hi GuiGuiGui

Could you please confirm which webserver are you using for this?

Thanks,
Kristiāns

Hello,

I'm using Apache/2.4.6 (Red Hat Enterprise Linux).

Please note I don't have the problem on firefox or Internet Explorer, only chrome 83.0.4103.97 at the moment

Hi GuiGuiGui

I'm not sure if this is actually a problem with Zabbix, seems like the browser itself has some memory leak issues. Have you tried other browsers based on Chromium (Brave, new Microsoft Edge)?

Thanks,
Kristiāns

Hello Kristians,

I have created this issue few months ago, then exchanged some messages with Edgars Melveris without having any relevant feedback. It looks like you're involved on this issue now. Could you give me the status about the first statement of this issue ? As I explained few months ago, the problem occurs on both Firefox and Google Chrome web browsers which let us think that the problem doesn't come from the web browser but from Zabbix.

Thanks for your feedback.

Hi,

I have the same problem as described above. I made some test and realized that the problem is gone if I remove the "Map Navigation Tree" Widget from my dashboard. I made some recording with the chrome developer tools. Here is the result with the Navigation Tree Map:

And here without:

I hope it can help to find the root of the problem.

Kind regards,
Fabien.

May be something worth to mention: we set the highest refreshing frequency (10 seconds).