https://rancher.com/ logo
Join Slack
Powered by
This message was deleted.
# general
a
This message was deleted.
b
f I install "Monitoring" it's going to install prometheus etc on one of the prod clusters.
Nope, it installs it in the local cluster only. There's weird UUIDs that pop up in namespaces that represent access to the downstream clusters for RBAC stuff, but there's not a OneMonitorToRuleThemAll as far as I can tell. Beyond that, the alerts you set up are namespace specific. So you'll end up having to create an AlertManagerConfig for each namespace and each cluster.
The upstream project supports a global notification, but it hasn't hit the released version of Rancher yet.
It'll be released... soon ™️
c
oh, thanks for correcting my misconception!
So if I just want to alert for nodes NotReady (hitting a weird Azure bug that is constantly locking them up), I may look for a simpler solution
b
I had a support ticket opened requesting they backport it to 2.7, but they haven't said if they're going to do that yet.
It depends on what cluster it's in.
And if there's already an alert generated for it or not.
I think that there's a generic alert in kubesystem for that by default.
c
sorry, what's kubesystem?
b
You'll just need the alertconfigs and secrets for right namespace it's firing in.
kube-system
namespace is one of the system namespaces that all (or almost all) k8s clusters have.
I can see if I can sanatize my script and gist it for you if you think it'd be helpful.
c
oh yeah, I thought 
kubesystem
might be part of Rancher or something.
b
It's part of Kubernetes in general.
Even OpenShift has it.
c
yeah I am aware of it, I was looking at it without the hyphen as something different.
b
oh yeah, sorry I guess my typing got lazy
c
appreciate you popping on this
b
np
Did you want that script?
c
I'll take any and all scripts lol
b
I think that's sane? You'll need to edit at least 3 lines
c
Much obliged
6 Views
Open in Slack
PreviousNext
Powered by