Hi all,

I have a patch in flight at the moment [0] to install check_mk server and 
compliment the already merged installation of check_mk agent [1] so my thoughts 
are now turning to how we would recommend adding new service checks.

The concept behind check_mk makes this really simple to do.  You just place a 
script that outputs "<status_code> <check_name> <performance_data> <message>" 
into the agent's "local" directory (/usr/lib/check_mk_agent/local on Ubuntu for 
example) and it will be picked up the next time an inventory of the system is 
run.

There are two ways that we can recommend doing this:

1) We ask users to update the check_mk_agent T-I-E every time they wish to add 
a new check
2) We ask users to distribute checks from their own T-I-E into the correct 
location

In my opinion, requiring an update to check_mk_agent for every new check is the 
wrong way of doing this as it means that all systems get all checks regardless 
of function.  Far more preferable would be option 2, however I'm open to other 
ideas, especially if they mean that organisations using this don't have to go 
through the review process if they have checks they wish to keep "behind the 
firewall" for IP/Licensing reasons.

Thanks in advance for any help people can give on this,

Matt

[0] https://review.openstack.org/#/c/87226/
[1] https://review.openstack.org/#/c/81485/


_______________________________________________
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Reply via email to