Similar to https://issues.redhat.com/browse/OSSM-5820
We copy the labels from the ingress of a knative service to the OCP route.
Among the labels that are used by ArgoCD is "app.kubernetes.io/instance:..."

It holds that: "Argo CD identifies resources it manages by setting the application instance label to the name of the managing Application on all resources that are managed (i.e. reconciled from Git). The default label used is the well-known label app.kubernetes.io/instance".

OCP route (unlike the ingress has no owner refs) and so it is considered as not being part of the resources deployed and is marked for pruning. Thus, the deployed ArgoCD app is out of sync (see attached files).

Reproduced with: https://github.com/openshift-knative/serverless-operator/compare/main...skonto:serverless-operator:routes?expand=1

Also verified that adding that label to a random resource eg. a pod, in the same ns where the app is creates an issue.