I am working for a customer who is running an enterprise-scale distributed application for high-end customers. It includes a crude self-made monitoring solution which is a pain to maintain so while looking for alternatives I came across LibreNMS which instantly caught my attention with its ease of adding new clients and its capable autodiscovery. It even reminded me of PRTG which is a really great commercial software on the surface but needs some ugly Windows server to run.
Alas, after some initial tinkering I have real trouble finding an angle trying to properly extend the built-in monitoring capabilities of libreNMS to include the set of custom values that need to be monitored on our installations.
At first I believed it would only be a matter of dropping in a bunch of check_mk compatible scripts but soon found out that for each of them I would have to hard-code php on the librenms side to have them displayed properly.
Next, I tried to add custom MIBs to the mibs directory and extend the snmpd config on our hosts, but here it also appears all OIDs must be hard coded into librenms. Also the trap capabilities of libreNMS appear to be pretty limited.
So finally, it looks like the only thing that can be used to dynamically extend librenms without codebase changes seems to be nagios scripts, but only as long as you do not want to combine graphs. Also the polling frequency of those scripts can not be set individually and the default is 60s which is a tad too coarse for some of my intended applications.
My question is, did I get all this right or did I overlook something obvious?
If libreNMS may be not the right tool for the task, are you aware of any other software that allows a similarly easy basic monitoring setup and more flexible extendability and might feel inclined to leave me a pointer?
Thank you very much in advance!