Debugging random crashes on Zabbix agent
Since a bit ago, the agent (the latest version available on Debian, 2.2.5 if i recall correctly) has been dying randomly on me for no reason, leaving nothing on the logs even with DebugLevel=4 (it's writing happily, then the next line is the typical "Starting Zabbix Agent" once i restart it). Looks like the agent dies after about between 15 and 30 minutes, randomly.
This is in multiple client machines, so i suspect there's either something on the server they don't like (why would something on the server kill the agent? no idea but it's common to all so i count it for now) or maybe one of the custom tests (which are plenty).
For now i have as a workaround a cron job to go every minute "is the agent running? if not, restart", but this is undesirable, i'd rather not have it dying on me all the time.
How could i debug it to see why is it dying, given than even debug mode says nothing? Apart from that, any ideas on why could it be dying or what should i check out?
This is in multiple client machines, so i suspect there's either something on the server they don't like (why would something on the server kill the agent? no idea but it's common to all so i count it for now) or maybe one of the custom tests (which are plenty).
For now i have as a workaround a cron job to go every minute "is the agent running? if not, restart", but this is undesirable, i'd rather not have it dying on me all the time.
How could i debug it to see why is it dying, given than even debug mode says nothing? Apart from that, any ideas on why could it be dying or what should i check out?
Comment