We have 254 nodeports currently in use (we have a number of environments sharing a cluster) and are using the default range. Is there a api parameter that I'm not seeing to bump this limit up?
You probably need to change the --service-cluster-ip-range= on the kube-api-server.