the cluster-reachable unit reconciles even if the cluster isn't reachable

I was looking at my broken workload cluster and noticed the following:

$ sylvactl watch Kustomization/my-capo-wc/sylva-units-status --once
...

 ✓ Kustomization/my-capo-wc/cluster-reachable - Resource is ready
⠈⠁ Kustomization/my-capo-wc/namespace-defs - ReconciliationFailed - Namespace/sylva-system dry-run failed: failed to get API group resources: unable to retrieve the complete list of server APIs: v1: Get "https://172.20.136.27:6443/api/v1?timeout=30s": dial tcp 172.20.136.27:6443: connect: no route to host

... even though the cluster is not reachable, the cluster-reachable unit happily reconciles.