[Grafana] Update Grafana dashboard

ray-project / kuberay

A toolkit to run Ray applications on Kubernetes

Apache License 2.0

982 stars 330 forks source link

[Grafana] Update Grafana dashboard #2106

Closed kevin85421 closed 2 months ago

kevin85421 commented 2 months ago

Why are these changes needed?

Upgrade the Ray Grafana dashboard from 2.5.0 to 2.9.0.
Add the Grafana dashboards for Ray Data / Ray Serve.

Related issue number

Checks

[ ] I've made sure the tests are passing.
Testing Strategy
- [ ] Unit tests
- [ ] Manual tests
- [ ] This PR is not tested :(

jjyao commented 2 months ago

What's the context of this PR?

kevin85421 commented 2 months ago

What's the context of this PR?

Currently, the Ray dashboard generates Grafana dashboard configurations at runtime. From my perspective, this is unusual behavior. Consequently, users need to copy the JSON configurations from the Ray head Pod and paste them into Grafana to create the dashboard. As a workaround, I also maintain a JSON configuration that users can directly use when following the Ray documentation.

kevin85421 commented 2 months ago

Convert it to draft. I will update the config to Ray 2.20 after I update the sample YAMLs to use Ray 2.20 images.

kevin85421 commented 2 months ago

Ray 2.20.0 has some issues, so I will not upgrade KubeRay to use Ray 2.20.0 by default. See https://ray-distributed.slack.com/archives/C02GFQ82JPM/p1714683469966249 for more details. Merge this PR.