I am using this version of Zabbix agent:
Zabbix is installed from official Zabbix repo.
I am trying monitor counts of certain processes on Ubuntu 24.04 servers using the proc.num check. For some reason the checks sometimes return 0 even tough the processes are running. Processes run as Systemd service and status confirms that they have been running when Zabbix reports that they did not.
The checks I have are like this:
Name of the user, application name and directory are macros in my template and values are set in the host level.
I have items for about 20 apps on the same server. Each app has one process running. Each check runs at 10 minute interval.
But as shown below these checks report quite often 0 processes running even tough all processes have been runnign without interruption.
No errors or warnings are logged into Zabbix agent's log.
The monitored applications do change their command line text.
This seems quite similar to this issue: https://support.zabbix.com/browse/ZBX-19614
But if undertsand correctly, that should be fixed in 7.0.
There is no particularly high load on the server. All others checks seem to be working normally as far as I can tell. I use checks from Linux by Zabbix agent active and some custom templates.
What could be causing this and how could I avoid it?
I know I could have Zabbix agent run a ps command or use external script and get results from those but I would prefer using standard checks for easier setup and maintenance.
Code:
zabbix_agent2 (Zabbix) 7.0.17 Revision 77c670937ef 23 July 2025, compilation time: Jul 23 2025 05:38:25, built with: go1.24.1 Plugin communication protocol version is 6.4.0
I am trying monitor counts of certain processes on Ubuntu 24.04 servers using the proc.num check. For some reason the checks sometimes return 0 even tough the processes are running. Processes run as Systemd service and status confirms that they have been running when Zabbix reports that they did not.
The checks I have are like this:
Code:
proc.num[python,app,,"^/apps/application_one.*?/apps/application_one/conf.ini"]
I have items for about 20 apps on the same server. Each app has one process running. Each check runs at 10 minute interval.
But as shown below these checks report quite often 0 processes running even tough all processes have been runnign without interruption.
No errors or warnings are logged into Zabbix agent's log.
The monitored applications do change their command line text.
This seems quite similar to this issue: https://support.zabbix.com/browse/ZBX-19614
But if undertsand correctly, that should be fixed in 7.0.
There is no particularly high load on the server. All others checks seem to be working normally as far as I can tell. I use checks from Linux by Zabbix agent active and some custom templates.
What could be causing this and how could I avoid it?
I know I could have Zabbix agent run a ps command or use external script and get results from those but I would prefer using standard checks for easier setup and maintenance.