Running arbitrary commands in escalation?
Hi. I want to execute an arbitrary command as part of escalations.
What i want to do is specifically, if the nodata() based trigger fires, first it tries to restart the agent on the remote host, before trying to contact anyone. It'd probably be a remote ssh command.
Since the problem at hand would be an unresponsive zabbix agent, trying to use the normal "remote command" method would fail by definition. The agent is not ok, who's gonna answer? and can the agent restart itself anyway?
I've thought about using a script alert that instead of trying to send some message to somebody, it'd ssh the host and restart the agent, but it feels hackish. I'll use that if i find no alternative, but i'd rather use something official if present.
Have i missed a feature in the docs that allows me to do this without hackish stuff? or do i simply go for the script?
What i want to do is specifically, if the nodata() based trigger fires, first it tries to restart the agent on the remote host, before trying to contact anyone. It'd probably be a remote ssh command.
Since the problem at hand would be an unresponsive zabbix agent, trying to use the normal "remote command" method would fail by definition. The agent is not ok, who's gonna answer? and can the agent restart itself anyway?
I've thought about using a script alert that instead of trying to send some message to somebody, it'd ssh the host and restart the agent, but it feels hackish. I'll use that if i find no alternative, but i'd rather use something official if present.
Have i missed a feature in the docs that allows me to do this without hackish stuff? or do i simply go for the script?