Getting Kubernetes monitoring in Nagios

[A3ch]

Admittedly this is not my strong suit. I have been trying to get kubernetes to report/monitor from within Nagios. I have not been successful. After running the command: juju deploy nrpe-external-master the following output shows with the sudo juju status command:

Model Controller Cloud/Region Version SLA Timestamp
controller test-juju localhost/localhost 2.5.0 unsupported 11:22:26-05:00

App Version Status Scale Charm Store Rev OS Notes
nrpe-external-master waiting 0 nrpe-external-master jujucharms 17 ubuntu

Machine State DNS Inst id Series AZ Message
0 started 10.181.20.35 juju-20b935-0 bionic Running

I can not get the nrpe app out of waiting status. When I tried running the command: juju configure nrpe-external-master nagios_master= [my external nagios XI install ip addy], I get an error that the command isnt valid. When I shorten configure ot config I get no output displayed but no error message.

Nothing changes with that current waiting status. The Kubernetes implementation I'm trying to monitor was already installed (not via juju) so any time I try to add a relation it basically says there is no kubernetes components to relate to? I have tried to figure out how to have juju see my separately installed kubes cluster and the furthest I got was getting the cloud to be seen.

All of this and I still dont see any kubes info in Nagios......I suspect until I figure out why the nrpe-external-master shows running it wont work.... I just dont know what do to to get past this point. I tried doing my googles but I have come up with nothing so far.

[scottwilkerson]

Unfortunately I cannot speak to the juju commands you are referencing because we are not the developers of that implementation nor do we have any documentation on it.

I scanned the Nagios exchange to see if I could find any plugins for monitoring kubernetes and only found one item shared
https://exchange.nagios.org/directory/P ... py/details

We do not have a setup that utilizes kubernetes in-house, however so I have not tested this.

[mcapra]

The Kubernetes implementation I'm trying to monitor was already installed (not via juju)

The official k8s documentation for monitoring, which is presumably what you're working off of, has this at the top:

This page assumes you have a working Juju deployed cluster.

I have a Chef managed k8s cluster that I can test this against, but I'm pretty sure it's not gonna work.

Have you explored spicing up your deployments with nrpe-server or a comparable image? What sorts of things are you trying to monitor?

[scottwilkerson]

thanks @mcapra