berngp
commented
8 years ago @kdkaergaard true, I have a (branch)[https://github.com/berngp/docker-zabbix/tree/feature/22-supervisord] that actually is meant to capture the effort to move to supervisord and avoid this. Adding syslog, as the link you mentions describes, is also a good idea.
My current problem is just that I have not been able to invest time in this project. Hopefully soon.
On a fairly generic setup, with some standard items, and some External checks, the container is slowly creating zombie processes. This seems to be related to the init problem, described in "A correct init process" section, on the baseimage-docker website, and in the pid-1-zombie-reaping-problem article.
Just for kicks, i've added a snippet of the Zabbix server, monitoring itself, with 3 hosts configured.