This is for large deployments in prod. Nomad and Consul servers at home are running on an old Rpi3 with 2G of ram managing several devices and about 10 vm's. Larger deployment of ~30vm's was managed with a nomad/consul cluster of 3x2cpu4ram with no issues.