Dashboard: Hi CPU usage on 1.10.1 Kubernetes Dashboard

[vasicvuk]

Environment

Installation method:
Kubernetes version: 1.11.7
Dashboard version: 1.10.1
Operating system: Windows 10
Browser: Google Chrome - latest version

Steps to reproduce

1. Go into Pods for namespace having 50 pods
2. Api is getting back results in 70ms
3. Front end acts very slow in rendering giving 100% of CPU usage

Here is the picture from Chrome task manager.

Machine specs

Intel Core i7-7500U
16 gb ddr4

just to know that this is not slow machine to work on in general

Expected result

To have faster load of pages

[floreks]

I am not sure how exactly your workflow looks like, but I don't see any issues with pod list view. Even having over 100 pods is not an issue.

PS. I am also using latest Chrome.

[vasicvuk]

@floreks Your screenshot shows the rendered page. Mine was taken during the rendering. During the rendering it uses from 60 to 190 CPU % on Chrome which quite a value.

Switch a little bit between pods and jobs and you will get the High CPU load.

Also the animations during the loading are very slow and the whole experience feels little blocked

[floreks]

I am not sure what to expect here. This is normal that to get something rendered ASAP, browser will allow all the resources available to render something. Such short spikes are normal and can be seen on most pages (gmail, github, etc.). Also the CPU value in Chrome Task Manager scales to your logical cores, so in your cases it would be 2 cores * 4 threads per core. Up to 800%. I have tested Dashboard on clusters with hundreds of resources and did not experience such slowdowns.

You could check if our head image, that is based on angular 7 and optimized will work better.

https://github.com/kubernetes/dashboard/blob/master/aio/deploy/recommended/kubernetes-dashboard-head.yaml

It can be accessed on: http://localhost:8001/api/v1/namespaces/kube-system/services/https:kubernetes-dashboard-head:/proxy/ if you are using kubectl proxy.

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[floreks]

/close

[k8s-ci-robot]

@floreks: Closing this issue.

In response to this:

/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[vasicvuk]

I tested Dashboard based on Angular 7 and it seems that it works better.

I really dont understand why the performance of current version is so bad specially when rendering Overview screen of some namespace.

[floreks]

New version is optimized better and rewritten from scratch. Some things work differently and it better handles async download of resources.