In order to create konflux component, we can either rely on https://gitlab.cee.redhat.com/releng/konflux-release-data/ (gitops approach) or apply things directly on the cluster.

The gitops approach is nice but it requires us to go through another set of merge-request, approval, … which can be slow, especially as we already store manifest on the git repository.

We should setup something in openshift-pipelines/hack that gets all manifests from .konflux in the different repositories and apply them to the cluster.

Bonus: we should handle "orphans", aka things that are no more defined (application gone, …)