topic Re: Maintenance windows and Openshift (maybe all k8s?) in Container platforms

Maintenance windows and Openshift (maybe all k8s?)

RuN3 — Thu, 01 Feb 2024 22:33:38 GMT

Hello!

We use the OLM operator with fullstack monitoring on openshift, but i am having alot of troubles getting maintenance windows to work.

For me the logical path would be to select "kubernetes cluster name" to define the entire cluster everything running on it.
But no services, nodes, anything cares about this.

I have had multiple chats with support but noone really seem to get this to work in a stable way for us.

Anyone have a tested and proven way to silence alerts during maintenance windows for openshift/kubernetes installations?

Re: Maintenance windows and Openshift (maybe all k8s?)

sivart_89 — Fri, 02 Feb 2024 02:16:21 GMT

Interested in this as well. I was just testing earlier today with the same thing. I had a problem created for not all pods ready within a workload of mine. I could not get the problem to be suppressed / show in a maintenance window even though the entire cluster was in one, at least I am assuming it was since I created the window. I don't know how to confirm that it is actually in a maintenance window like you could would a standard host. Normally for them you could see a grey bar on the host and it show the maint window tied to it.

Re: Maintenance windows and Openshift (maybe all k8s?)

apanoobee — Tue, 19 Mar 2024 10:05:30 GMT

@RuN3 @sivart_89 Would be interested to know if you guys managed to find a solution for this as we're also experiencing the same issue.

Re: Maintenance windows and Openshift (maybe all k8s?)

sivart_89 — Tue, 19 Mar 2024 21:33:42 GMT

I recently found I believe what is working. From what we were seeing just using the entity filter for the cluster was not working (see below) but when using the mgmt zone filter and choosing the mgmt zone that is specific to that 1 cluster, it seemed to work.

You can create a mgmt zone from the Kubernetes screen and clicking the ellipses. There may be some fancy stuff you can do with tags but that was much more complex and we already had a mgmt zone for each of our clusters.