The majority of information out there, including tutorials and blog articles about others' successful deployments, comes in the form of very high-level overviews. Everything I've found is an introduction to getting a basic docker instance running. There is very little useful information out there as to how to run a proper multi-host cluster.

There is core Docker. Tack on docker-machine, docker-compose, Swarm, and the dozens of 3rd-party cluster management abstractions such as Rancher - and the intensity of the headache never stops growing.

It sounds wonderful, but there is so much to learn to be able to tackle a full production stack. It's one thing to successfully launch a working cluster after hours of manual tinkering. It's a separate beast altogether to fully automate setting up a new cluster by issuing a single command, taking into account consistent configuration of: secure networking, persistent volumes with backups, deployment of container configuration and VCS codebases (ex: nginx vhosts and your code itself), etc.

My goal is to set up an entire project in such a way that there is a single suite of automation that can deploy all environments: development VM, staging, and production.