Monitoring remote services (new to zabbix)

[lsimpson]

I have a working installation of Zabbix 3 on Centos 7.2.

The monitored hosts have all been auto-discovered, and I am getting good graphs of CPU and Memory use.

I would now like to monitor the state of a service on the remote hosts - sssd - but I am unsure where to start.

Is this an External Check or do I need to put it into an SSH Check?

I know how to eyeball the data locally - on any server run

systemctl status sssd

And I know that I want to run something more like:

`systemctl is-active $SERVICE` != "active"

on the host in question. But how do I do that (or "what is the best way to do that?").

I should note - in case this affects the answer - that once monitored, I would also like to add triggers: if it enters the state failed, email me and then attempt to restart it locally.

[guzzijason]

You have options. The SSH Check you mention is meant to use the SSH protocol to perform agentless checks of arbitrary items on the target server. That may not necessarily be what you are looking for.

You can perform a simple check, like so:

net.tcp.service[ssh]

https://www.zabbix.com/documentation/3.0/manual/config/items/itemtypes/simple_checks

Or, you could use an agent check to count the number of sshd processes running on a host:

proc.num[sshd]

https://www.zabbix.com/documentation/3.0/manual/config/items/itemtypes/zabbix_agent

__Jason

[lsimpson]

Thanks @guzzijason - I don't think that's quite what I'm after. The service I wanted to monitor, in particular, is sssd (FreeIPA authentication module) not sshd (secure shell).

While I could poll sssd via tcp, but I'd rather check on the process itself.

[lsimpson]

In some ways what I'm asking is "how can I monitor all my services on hosts". They all have agents. Can I just watch systemctl for status changes?

<edit>

Ok, I've found this post on the zabbix feature requests page. I have created a file on one of my hosts called /etc/zabbix/zabbix_agentd.d/userparameter_sssd.conf

With the content

 # checks to determine if specified unit is active, failed or enabled
 UserParameter=systemd.unit.is-active[*],systemctl is-active --quiet '$1' && echo 1 || echo 0
 UserParameter=systemd.unit.is-failed[*],systemctl is-failed --quiet '$1' && echo 1 || echo 0
 UserParameter=systemd.unit.is-enabled[*],systemctl is-enabled --quiet '$1' && echo 1 || echo 0

but still don't quite understand what I'm doing - should this file be on the host or on the zabbix server? Also, now that it exists, how do I use it? (I will do more reading.)

[guzzijason]

Sorry I misread 'sssd' as 'sshd'. However, if 'sssd' is the actual process name, this should still give you something: proc.num[sssd]. That will give you the number of sssd processes running on the host.

If you want to go the userParameter route, you would install the file you mentioned on your target host, make sure your zabbix-agentd.conf is including the files in zabbix_agentd.d directory, and then restart zabbix-agent. Then, in the item configuration for this host (on the server), you would set up a zabbix agent type check for systemd.unit.is-active[sssd] (and/or one of the other keys specified in your custom userparameters.