fix: avoid hard-code Prometheus datasource id

ray-project / kuberay

A toolkit to run Ray applications on Kubernetes

Apache License 2.0

990 stars 330 forks source link

fix: avoid hard-code Prometheus datasource id #2043

Open hustshawn opened 3 months ago

hustshawn commented 3 months ago

Why are these changes needed?

Currently the kuberay-api-server and kuberay-controller-runtime Grafana dashboard hard coded the Prometheus data source id, which is not re-usable by people. Fixed the hard code.

Related issue number

Checks

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

kevin85421 commented 3 months ago

Would you mind providing more details about how you test this PR manually? These two files are not mentioned in the Ray documentation. In addition, these files haven't been updated for two years, so I am unsure whether they work or not. I need to manually test this PR.

hustshawn commented 3 months ago

Thanks @kevin85421 . Steps include,

Setup the Prometheus and Grafana. follow this doc
In Grafana, import the dashboard as the snapshot. Copy and past the Json mentioned and import

kevin85421 commented 3 months ago

cc @Yicheng-Lu-llll would you mind reviewing this PR?

kevin85421 commented 3 months ago

Setup the Prometheus and Grafana. follow this doc

The document doesn't include any service monitor or Pod monitor for the KubeRay operator Pod. Is Prometheus able to fetch the metrics?