so according to latest messages in channel there are no such functionality and I should use native module instead

the problem is - I can't convert Output types onto simple types to use it with other modules

And say you didn't use Pulumi but did everything manually through

kubectl

, what would "finding the IP and using it" look like?

Getting information into Pulumi is relatively straightforward: You query the information via the Kubernetes SDK within your program:

def find_deployment_by_labels(labels: dict[str, str]) -> (str, str):
   """Utility function to find an existing deployment."""
   ...
   return (namespace, name)

dep_namespace, dep_name = find_deployment_by_labels({"internal/app": "my-app", "internal/team": "my-team"})

my_new_secret = k8s.core.v1.Secret("my-new-secret", 
                                   metadata={"namespace": dep_namespace,
                                             "name": "my-shiny-new-secret"}
                                   )

Alternatively, you can pass in information as configuration but if you can query it just when the program is running, that's preferable in all cases, I think.

Getting information back to your team can work in several ways: You can do it within the context of Kubernetes, either by having them find resources through labels etc. or by maintaining a ConfigMap. Or you generate stack outputs and pass it to them outside of Kubernetes. Again, doing things within Kubernetes should be your best option.

I think it might help to think about Pulumi as "just" one way to create and evolve Kubernetes resources, similar to (but more powerful than) kubectl or Helm.