When implementing the OpenStack controller for cluster-capi-operator (CCAPIO), we opted to do so as a wholly separate controller in the OpenShift fork of cluster-api-provider-openstack (CAPO). While there were good reasons for doing so, no other provider is doing this and it leaves us maintaining a snowflake configuration and results in circular dependencies. We should correct course and migrate the controller into CCAPIO, dropping the implementation currently found in CAPO in the process.