kubegraf
kubegraf copied to clipboard
devopsprodigy
Applications Overview page partly broken since Kubernetes version 1.18.X
Hi,
we're using kubegraf version 1.4.2. It was working fine but since we've upgraded Kubernetes to version 1.18.X, the Applications Overview page is partly broken. The controller-manager and scheduler are marked as down but both are running fine. The error messages are Get http://127.0.0.1:10251/healthz: dial tcp 127.0.0.1:10251: connect: connection refused/Get http://127.0.0.1:10252/healthz: dial tcp 127.0.0.1:10252: connect: connection refused
The reason for that is that the componentstatus view is deprecated (as described here) and the urls for the mentioned components have changed to secure ports.
controller-manager (see here):
- old url: http://127.0.0.1:10252/healthz
- new url: https://127.0.0.1:10257/healthz
scheduler (see here):
- old url: http://127.0.0.1:10251/healthz
- new url: http://127.0.0.1:10259/healthz
Would be great if this could be fixed.
Thank you.
I have now the same problem with our first kubeadm set up HA-cluster (else we only have rancher-clusters atm where it works - you can see they use "--address 0.0.0.0" in their configs) - as far as I could find out those scheduler and controller-manager ports are only useful for those health-checks and default only are listening to 127.0.0.1 (with newer kubernetes versions) on the nodes and there is no real need to configure them to listen on any other interface as far as I understood - only if you you separate them from the nodes where the kubernetes API-servers are running.
so I think in this case we can only ignore the state - or reconfigure them for the moment with deprecated "--address" and "--port" settings which might be not a good idea.
maybe there should be an option where you can select if you want to see the state of the scheduler and controller-manager in the Applications Overview or not
