workload clusters: fix generation of sylva-units-status configmap

Relates to #1831 (closed)

Needed for !3101 (merged)

This works as expected:

https://gitlab.com/sylva-projects/sylva-core/-/jobs/8276560285/artifacts/external_file/management-cluster-dump/ConfigMaps.summary.txt

kubeadm-capm3-virt                                                sylva-units-status                                                 5      3m23s
...
sylva-system                                                      sylva-units-status                                                 5      27m
...

⚠ Note well that unfortunately 😭 this change is not sufficient to make _internal.is_upgrade work in the case of an upgrade from a previous Sylva version, because in that case the starting point has the ConfigMap in the wrong cluster in the wrong namespace.