Because we can't begin to agree on how to do this in a standard way. The whole idea of git oriented flows seems to be an afterthought rather than being considered an important part of a standard production k8s deployment from the start. Here are some competing options I see after a quick glance in this thread:

  - kubectl apply -k ./deploy --prune
  - heml install
  - helm template + kubectl apply + kubectl diff
  - Kustomize
  - Argo CD
  - Octopus
  - JSONNet
  - Cue
  - Pulumi

And because this is an advanced topic (most of the above require a large investment reading docs and github issues), I'm guessing many/most people never get beyond ad-hoc kubectl commands and letting their cluster run wild.