SweetOps #kubernetes for November, 2023

Archive: https://archive.sweetops.com/kubernetes/

2023-11-03

2023-11-05

2023-11-08

Andy

01:58:31 PM

https://eksclustergames.com/

EKS Cluster Games

The mission? To identify common AWS EKS security issues and vulnerabilities and learn how to exploit them in practice.

venkata.mutyala

03:03:59 AM

I’m using kube prometheus stack and it works well but today I found that I can’t find a dashboard that tells me my overall capacity usage. Ex. if my cluster has 4 cpu total and i’ve requested 3 cpu already leaving 1 cpu left to be requested. Does anyone have a recommended community dashboard for this? Also, any other dashboards you folks would suggest ontop of the default kube-prometheus-stack installation?

venkata.mutyala

03:06:38 AM

Alternatively, i’ll probably build one and just publish it as a community dashboard. Any other high level metrics you folks would find helpful?

venkata.mutyala

03:08:10 AM

I think i may have answered my own question: https://grafana.com/grafana/dashboards/?search=capacity

Dashboards | Grafana Labs attachment image

Browse a library of official and community-built dashboards.

venkata.mutyala

03:24:59 AM

Upvote please: https://github.com/prometheus-community/helm-charts/issues/3977

#3977 [kube-prometheus-stack] Adding a capacity dashboard

Is your feature request related to a problem ?

The current dashboards that ship with kube prometheus stack do not show capacity.

Describe the solution you’d like.

I think adding a capacity dashboard would be very helpful to users of kube-prometheus-stack. Here are a couple of examples I was thinking of implementing:

https://grafana.com/grafana/dashboards/13125-kubernetes-capacity-planning-limits/
https://grafana.com/grafana/dashboards/16237-cluster-capacity/

Describe alternatives you’ve considered.

I can build a dashboard for myself but I think this would be a good addition to the helm chart so that everyone can benefit from it.

Additional context.

I’d be happy to make a PR myself or let someone pick up this idea. I’m hoping to just get this approved and get additional feedback before I go off and build it for kube-prometheus-stack