Infrastructure discovery with etcd

Infrastructure discovery with consul is easy, mostly due to its known semantics of a server, a service and associated health checks. But using consul in production is a topic on its own – scale a running cluster is hard, ensuring all the zombie’s are gone need to be implemented on your own and the solution for initial cluster startup (atlas) is commercial. With etcd you can achieve easily the same infrastructure discovery concept as consul has, with just a couple of lines of code. This post will guide you through a good alternative of infrastructure discovery, if you don’t want to use consul in production.