[FEATURE REQ] Kubernetes Monitoring and Dashboard

[anirudhRowjee]

Is your feature request related to a problem? Please describe.
Kubernetes Monitoring and Dashboarding

Describe the solution you'd like
I believe a lot of the existing architecture of Grofer can be re-used to show data relevant to Kubernetes. A simple dashboard can be implemented, with support for features including (but not limited to)
0. Listing all available pods, nodes, and similar data sources made available by the Kubernetes API

1. Pod-Specific Actions on keypress (a la grofer proc)
2. Charts for available metrics
3. In-Memory time-series storage of data collected (could perhaps make use of some of the existing capabilities of the export feature)

Describe alternatives you've considered if any
The tool of choice for this appears to be k9s, but I believe the learning outcome of implementing this here would be worth the effort.

Additional context
None.

I would love to work on this feature if it's approved! Thank you for the time.

[Samyak2]

Sounds like a great idea! Do you want to work on it?

@MadhavJivrajani thoughts?

[MadhavJivrajani]

Huge +1
Happy to help in any way possible :)

[anirudhRowjee]

I will! Thank you for the opportunity :D

[anirudhRowjee]

I'm considering using the go-client library to scrape kubernetes metrics, and limiting support to running grofer from outside the cluster, so as to have a simpler baseline goal to approach.

I'd love to hear the maintainers' feedback on this! 😄

[anirudhRowjee]

I'd also love to add some documentation on navigating the codebase and a document explaining the new architecture! DO let me know if you want me to open a new issue / PR just for this.

[MadhavJivrajani]

I'm considering using the go-client library to scrape kubernetes metrics, and limiting support to running grofer from outside the cluster, so as to have a simpler baseline goal to approach.

I'd love to hear the maintainers' feedback on this! smile

@anirudhRowjee you have the right idea. client-go can be used to interact with the Kubernetes API. However, for the purposes of scraping metrics, there are a few things to keep in mind:

• The cluster needs to have metrics-server deployed, only then will you be able to scrape metrics using the Kubernetes API.
  • There are limitations on the frequency of scraping metrics that need to be considered during implementation.
• For the specific use-case of scraping metrics, there is a metrics client available that you can check out.

client-go is the perfect choice to get the available resources in the cluster such as the Pods, deployments etc.

[MadhavJivrajani]

I'd also love to add some documentation on navigating the codebase and a document explaining the new architecture! DO let me know if you want me to open a new issue / PR just for this.

Absolutely! Go for it. Feel free to open a new issue about this.

[MadhavJivrajani]

@anirudhRowjee fyi - https://github.com/vladimirvivien/ktop