search

digitalocean / DOKS

Managed Kubernetes designed for simple and cost effective container orchestration.
https://www.digitalocean.com/products/kubernetes/
Apache License 2.0
82 stars 5 forks source link

No metrics for cluster in version 1.20.2-do.0 #24

Closed chevdor closed 3 years ago

chevdor commented 3 years ago

Hello,

I am fighting to deploy the metrics-server and see CPU/Mem usage in the dashboard using a control plane with 1.20.2-do.0.

kubectl top pods returns:

Error from server (ServiceUnavailable): the server is currently unable to handle the request (get pods.metrics.k8s.io)

I have tried many things and current everything *appears * to be running properly on the cluster, but still no metrics.

timoreimann commented 3 years ago

@chevdor if this is a vanilla installation of metrics-server, then you probably need to set --kubelet-preferred-address-types=InternalIP. (See also #2 and especially https://github.com/digitalocean/DOKS/issues/2#issuecomment-530822912.)

One easy way to ensure the metrics-server is working correctly on DOKS is to install it from the DO Marketplace. I just verified that it works correctly on DOKS 1.20.

chevdor commented 3 years ago

This is what I initially did and the dashboard works indeed, but not the charts/metrics.

On Mon, 15 Feb 2021, 12:21 Timo Reimann, notifications@github.com wrote:

@chevdor https://github.com/chevdor if this is a vanilla installation of metrics-server, then you probably need to set --kubelet-preferred-address-types=InternalIP. (See also #2 https://github.com/digitalocean/DOKS/issues/2 and especially #2 (comment) https://github.com/digitalocean/DOKS/issues/2#issuecomment-530822912.)

One easy way to ensure the metrics-server is working correctly on DOKS is to install it from the DO Marketplace https://marketplace.digitalocean.com/apps/kubernetes-metrics-server. I just verified that it works correctly on DOKS 1.20.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/digitalocean/DOKS/issues/24#issuecomment-779155998, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAFULJDGHMAOOVKJP3NPA73S7D7UHANCNFSM4XTI3GYA .

timoreimann commented 3 years ago

@chevdor by charts/metrics, you mean Helm charts?

Can you share any manifests describing your configuration?

chevdor commented 3 years ago

No I mean the metrics charts in the dashboard, showing cpu and men usage over time, for nodes and pods.

Although, after chatting with DO support, I have been told that it cannot (unfortunately) work because DO does not deploy the scrapper on the DO dashboard. This is also a dashboard users cannot modify.

So the only workaround atm is for users to install the dashboard again, giving them a second version they can customize.

timoreimann commented 3 years ago

That's (unfortunately) correct for the time being. You can follow #21 to be notified when we support metrics in the dashboard.

timoreimann commented 3 years ago

Let's close this ticket to keep any dashboard-specific discussions around metrics focused in one issue.