[Serve] Basic serve deployment example in documentation throws error with Python 3.11

Ray Libraries (Data, Train, Tune, Serve)
1

I have raised a github issue on this (#36319), but thought I would ask in this forum to get a better understanding of what is meant by experimental support for Python 3.11.

My observation is that the basic deployment example in “Key Concepts” of the Serve documentation throws an error when run with Python 3.11.
https://docs.ray.io/en/latest/serve/key-concepts.html#serve-key-concepts-deployment

@serve.deployment
class MyFirstDeployment:
  # Take the message to return as an argument to the constructor.
  def __init__(self, msg):
      self.msg = msg

  def __call__(self):
      return self.msg

my_first_deployment = MyFirstDeployment.bind("Hello world!")
handle = serve.run(my_first_deployment)

All details are in this issue (#36319). The error seems to point to a protobuf error, but not sure if there is something else happening in the background.

ValueError: Enum type ray.serve.DeploymentStatus: unknown label "DEPLOYMENT_STATUS_DeploymentStatus.UPDATING"

Is there an expected Ray release that will support 3.11?
And would it be possible to further elaborate what is meant by experimental support?

2

@jharaldson Experimental support means that we are releasing to the community to try and give us feedback, and we may change the API or interface before going to GA. As such, we don’t recommend using it in production for only testing and experimenting.

Python 3.11 in Ray 2.5 is still experimental.

cc: @rickyyx @eoakes @Sihan_Wang

3

Thank you for quick answer!
Any type of workaround would be of interest, also to know if there are potentially more issues to expect.