Hi!
I am using this Helm Chart to deploy a Ray Cluster on KubeRay (that I am using to train ML models). I now want to use the same cluster to serve some applications. I’ve added a GRPC ingress for the client which allows me to start serving such applications. However, my requests cannot reach the served applications. As far as I understand, I’d need to set up a RayService and an ingress for the port 8000 that binds to the RayService. How do I create this RayService CR and make it bind to my existing Ray Cluster / how do I add this RayService to the Helm Chart? Does a RayService always create its own cluster? If so, can I use it similar to the cluster that the Helm Chart creates? I’ve already worked a bit with the Ray Serve examples for KubeRay (like this one) but all examples that I am aware would create their own cluster.
Thanks for the help!