Support existing management cluster

For a commercial use case where a production management cluster is already deployed, there's no need to perform bootstrapping and pivot. In this case, we only need to run apply.sh with a customized values.yaml to ensure the required Sylva units are working in harmony with the existing cluster.

To make this work, we need to remove the following constraints:

1. management-cluster-kubeconfig should not be mandatory as sylvactl is capable of working with the default kubeconfig (in ~/.kube/config).
2. cluster unit should not be mandatory. This essentially renders the enabled flag in the cluster unit ineffective. We should let the unit's enabled flag to determine whether the given unit should be enabled, for consistency.