Executing monitoring scripts on remote device

Help
1

Hi,

I’ve been using the Nagios services for ping and http testing and initially I was assuming given a howto I found on the net that the associated host you select when you define the service was the one executing the plugin, if installed on it, but based upon an article in this forum I am not so sure anymore. The funny thing is that I do get different results. Can anyone verify whether all is executed locally on the LibreNMS host or if the associated host is the one executing the test Nagios plugin ?

I looked into nrpe but it seems it cannot take variables so I would have to hardcode the tests on each client and not be able to define them in LibreNMS.

What I am looking for is to run some ping and http tests from devices in various remote locations and then get the data back to the LibreNMS server and later consolidate the results into a combined Grafana overview.

If none of the plugins or nrpe will do this in an easy way, are there any other best practises around for this, such as using SSH or similar ?

Regards,

Jesper

2

They are all excited from the local host (Librenms)

I have heard some users using NRPE to do this on remote host.

3

OK, so I looked a bit more into it and found that NRPE works when configured properly but is deliberately limiting what you can execute remotely to make it safer.

Other option is to use check_by_ssh which allow more flexibility (and potentially more issues).

I’ll try both and see if the output generates a RRD. I’ll have to format the SSH output to the same as Nagios scripts I assume.

Regards,
Jesper

4

Hi @dassic / Jesper,

what was the result of your evaluation if I may ask?
I am looking to perform some checks via ssh on ESX boxes and wonder how to get the command results into a graphable format …

Thanks & regards

Edit: I found out how to use check_by_ssh with graphs, that will be ok, thanks :slight_smile:

5