Orchestration and the On-Prem Split

Not written yet. Questions to answer here: what problem Kubernetes solves that running processes on machines does not, what the control plane actually does, what a reconciliation loop is and why the whole design is built on one, how storage and networking attach, what an operator is, and then the split: which workloads belong in a cloud region and which have to run on-premise, what running your own hardware costs in operational terms, and how one system spans both.