In my company we are on a custom vendor software built on top of openshift, which exposes a python api to start and stop pods.
I am able to manually create Ray clusters by calling something like cloud.launch_workers(n, resources, code="ray start - -address <Ray head>
Would i be able to get my Ray cluster to auto scale by telling Ray to call this command when it wants more nodes? I looked into node provider but it wants a lot more info which I can’t give (e.g. Capability to ssh into these pods)
Thanks, much appreciated!