I'm interested in using collectd as a replacement to remotely executing
nagios plugins on client machines to gather data.
I originally intended to use collectd-nagios to pull data from collectd
through the unixsock, but then I saw this post:
http://tokkee.org/talks/collectd-NEB.pdf
And thought that it might be pretty simple to write a plugin to receive
threshold notifications and submit them to the nagios external command
queue. Nagios just needs to be configured for passive checks, and to
process it's command queue regularly.
The original post I think envisioned a more deeply integrated nagios
event broker, but the command queue hack *should* work(?)
To my thinking this approach should capture the strengths of both nagios
and collectd. It uses nagios's well-developed dashboard, email
notification system, and dependency tracking. And it leverages
collectd's strength in gathering data. Finally it should eliminate the
need to run nrpe on clients.
If no one is working on this I'll hack together a perl plugin and post
it back.
RP
_______________________________________________
collectd mailing list
[email protected]
http://mailman.verplant.org/listinfo/collectd
