Before I knew it, crond was dead ...
It's unclear if it was killed by the Linux OOM Killer or if something else was the cause. As a result, let's encrypt was not automatically renewed and the certificate expired.
This is fine.
I have to warn you if crond dies in zabbix!
.... I don't have any items ... orz
You should make it, you should make it ~.
For the time being, create a new template Crond with zabbix server. Put crond on the item with key proc.num [crond].
This should be roughly okay, but in rare cases Zabbix-agent that proc.num [crond] is not taken ...
In such a case, the agent side notifies the server of life and death. You can check the life and death of crond with status. /etc/rc.d/init.d/crond status
It is normal if there is the word running. If not, it is abnormal. You can check by counting running.
/etc/rc.d/init.d/crond status|grep -e running |wc -l
The result is output from zabbix-agent. zabbix_agentd.conf UserParameter=crond,/etc/rc.d/init.d/crond status|grep -e running |wc -l
Add crond to the item on Zabbix server side and set the trigger.
Should I add it to dozens of zabbix-agents that I'm monitoring? Shit!