You have a computer. That computer does exactly one thing, and anyone can make a copy of it, and tweak it slightly to fit their needs. That’s a Docker container. It needs to run on other computers.

You need a way to manage multiple Docker containers, and the networking between them, and the items you need to customize them. You also want to make sure that they aren’t going to bump into each other accidentally, so you need to sequester them in different areas. You also need to make sure that only certain people have access to those different areas, so you set up authentication in a fashion that reflects that. You also need to manage the computers the Docker containers are running on, to make sure that only certain hosts run certain workloads, and that there’s a difference between controller and worker nodes, and that the networking is set up correctly between them, using a custom thing called a container network interface, of which there’re multiple choices and you need to decide which best fits you needs. You also need to make sure that you have a permanent storage solution that the Docker containers can use, so that if you lose a worker node you don’t lose all the data for all the containers running on it. All of this is run through API calls to the controller nodes. That’s what I remember off the top of my head for Kubernetes, and I think I’m forgetting stuff.

We aren’t crying because of you, we did this to ourselves.