Zabbix Agent

[danlogu]

Hello, experts!

I have been investigating an issue that arose on our network a few days ago. We use a proxy for monitoring, but it has been showing the following alert:

Linux: Zabbix agent is not available (for 3m)

And we don't know what may be causing it. The weird thing is that communication between both is ok, our proxy is still collecting information and sending it to our server in real time. We tought it may be related to the time sync since the proxy was around 3 minutes ahead, but when solving it the alert didn't solve. Do you have any idea what may be causing this?

[cyber]

Trigger config?
Active or passive agent?

[danlogu]

Trigger is the default for template Linux by zabbix agent: max(/Linux by Zabbix agent/zabbix[host,agent,available],{$AGENT.TIMEOUT})=0
Active mode.

It is not the configuration files inside /etc/zabbix/ as they are both pointing to our server. Tho i haven't checked if someone made a firewall change that could have blocked port 10050.

[danlogu]

Did a little investigation. It is not a firewall rule.

[jhboricua]

Trigger is the default for template Linux by zabbix agent: max(/Linux by Zabbix agent/zabbix[host,agent,available],{$AGENT.TIMEOUT})=0
Active mode.

It is not the configuration files inside /etc/zabbix/ as they are both pointing to our server. Tho i haven't checked if someone made a firewall change that could have blocked port 10050.

You're using passive checks, not active. The template "Linux by zabbix agent" uses passive checks, and the trigger you mentioned is in it. The template that uses active checks is "Linux by zabbix agent active" and the corresponding trigger expression is slightly different.

[danlogu]

You're using passive checks, not active. The template "Linux by zabbix agent" uses passive checks, and the trigger you mentioned is in it. The template that uses active checks is "Linux by zabbix agent active" and the corresponding trigger expression is slightly different.

Is it? I'm no expert in Zabbix, but in the Proxies section I can see that it is in Active mode. Or are those two different things?

[Brambo]

Is it? I'm no expert in Zabbix, but in the Proxies section I can see that it is in Active mode. Or are those two different things?

First of all, proxy are in simple terms a stripped down server (so no frontend, no trigger evaulations etc).
Proxy can only be in Active mode OR passive, not both at the same time.

Agents can be in both at the same time, but that depends what you configure in the conf file. Like does the agent accepts passive check from the origin IP. Similar for active checks where does your agent try to get it's config from.

[jhboricua]

Proxy mode refers to communication between the Proxy and Zabbix server for the purposes of sending the metrics that the proxy is collecting from agents pointing to it. As Brambo indicated, it can only be one or the other.

For monitoring the proxy metrics you can use the Zabbix Proxy Health template, which uses internal checks and is not dependent of an agent. If you also want to monitor the Os health of the proxy you'd need an agent installed and the proper Linux template. The trigger you mentioned at the beginning belongs to the 'Linux by Zabbix Agent' template, which uses passive checks. I'm only saying that because the active version of that trigger is:

nodata(/Linux by Zabbix agent active/agent.ping,{$AGENT.NODATA_TIMEOUT})=1

Which template(s) did you assign to the proxy for monitoring purposes?
Is there an agent installed on the proxy and if so, where is the agent pointing to, itself or the zabbix server (Server and ServerActive conf lines on agent)?