Sometimes linux systems are under memory pressure due to unexpected behavior of software. In this situation the builtin ooom killer of the linux kernel kills processes to resolve the memory pressure situation.
To monitor this is important, because systemd and upstart simply restarts killed processes and sometimes the operators of the system were not aware that the problem occurred.

It would be nice to have zabbix items which provide information about the behavior of the oom killer - i.e.:

• how many ooms where invoked which zabbix agent is running?
• the commandline and pid of the killed process
• what was the reason for the oom killer? (if it is easy to have a low number of concrete situations)

Probably this can be a active item, which is sent directly when the situation appears.
To implement this, it might be an option to to use the event interface described at https://www.kernel.org/doc/html/v4.12/media/uapi/v4l/dev-event.html