It's a dedicated server. I originally went for a 64G server to manage them all, but there was a limitation in the number of IPs I could get, so now I have a 64G and a 32G. I could have easily done the nodes I wanted with just two 32G servers. It is a potential point of failure, but we're already up over 5k nodes, so I don't think it's a real concern for the network. I'll be adding even more monitoring scripts so it'll text me if needed, as I've done with my witness servers.