Support patching from common data sources

[wcarlsen]

Use case 1) It seems relevant for us to abstract away "sensitive" inputs e.g. kafka cluster id's and environment id's for our end users, so that we can map variables like environment (enum: ["prod", "dev"]) to what the resource actually require. This means that the end user do not need to care about cluster id's etc.

2) This makes it also rather hard to distribute or share compositions without at least the artifact (docker image) being private and hence not shareable.

See https://github.com/crossplane/crossplane/issues/2099

[SEQUOIIA]

Screenshot showing what we are trying to achieve. (all "unique" data is generated and not actual endpoints/ids we use ourselves)

[samidbb]

Test this solution https://github.com/crossplane-contrib/provider-kubernetes/blob/main/examples/object/references/patches-from-resource.yaml requires provider-kubernetes v0.3.0

[samidbb]

currently solved by the solution that uses provider-kubernetes