I asked basically this same question a few days ago, you can see the answer I was given here.
I am successfully sending alerts from the the same workspace to different endpoints based on workspace name and tags as follows:
data_stream.namespace:"my_namespace" AND tags:("my_app" AND "my_env")
The one issue I'm having is that the uptime rules don't appear to respect the alert.status.enabled switch in my tests. I'm planning on resolving this in the short term by ignoring anything with a disable_alerts tag set, but I'm hoping there's a better solution.