lack of important checks in zabbix.

[leoh]

1 - How to monitor CPU utilization (vmstat or sar comand in Solaris). It's not cpu_average (uptime).

2 - How to monitor a process that is not listed in zabbix_agentd comands? Like check_service[syslogd] ? Does Zabbix have a comand to check if the daemon is alive? This should work for everything. Of course it would be just a simple ps|grep without port checking or so.

3 - What about monitoring logs? Any idea of when will it be released?

[James Wells]

Greetings,

From reading the title of your thread, the first thought was that you were stating thst Zabbix was missing things, but when I read your actual thread I realize that you either mis-typed the title, or that you had never even read about Zabbix. I really hope it was the former.

1 - How to monitor CPU utilization (vmstat or sar comand in Solaris). It's not cpu_average (uptime).

Yes, Zabbix does this fairly well right now and the instructions are fairly straight forward. Simply create an item that checks whichever of the system.cpu.<foo> items you want to check. For detailed list of checks, I would suggest;
The Zabbix Manual -- Items Section

2 - How to monitor a process that is not listed in zabbix_agentd comands? Like check_service[syslogd] ? Does Zabbix have a comand to check if the daemon is alive? This should work for everything. Of course it would be just a simple ps|grep without port checking or so.

Yup. Zabbix does this fairly well as well. You simply create a custom item to check for the service you want, or you can use many of the ones already built into the system. I would suggest reading;
The Zabbix Manual -- Cookbook Section and the Zabbix Cookbook Forums

3 - What about monitoring logs? Any idea of when will it be released?

Hmm, this is a tough one as Zabbix has supported this feature for a while now.

[leoh]

Regarding item 1 and 2, thank you.

Regarding item 3, are you being sarcastic? Zabbix has or has not text logs monitoring facility? Windows event log I know that it has. What about /var/log/messages, for example?

I read the manual and it said "Not yet implemented". Or something like that.

[Alexei]

Log file monitoring has already been implemented. Not sure if it is properly documented though.

Monitoring of log files:

log[filename]

Make sure that type of the check is "Log".

[gatisa]

should it work with 1.1beta2?

here is 1.1beta2 agent log:
018621:20051021:104731 Got line:log[/var/log/messages]
018621:20051021:104731 Sending back:ZBX_NOTSUPPORTED

[Alexei]

Define it as Active Check to make it work.

[Nate Bell]

Yes, Zabbix does this fairly well right now and the instructions are fairly straight forward. Simply create an item that checks whichever of the system.cpu.<foo> items you want to check. For detailed list of checks, I would suggest;
The Zabbix Manual -- Items Section

I've tried this with several versions of Zabbix (I'm currently using 1.1beta1) and have not had any luck with it. It reads in a value just fine, but the value never changes. I used the system.cpu.system1 nice1 user1 and idle1. Any thoughts on this? Currently, I'm using a script that reads these values from top and returns them. Unfortunately, top is pretty slow, making the script pretty inefficient.

Nate