Surviving AI with Minimal Viable Kubernetes
AI is changing what it means to be a software developer, founder, and technologist. It gives individuals the ability to work across disciplines—but that freedom also creates a new problem: how do you build and run the infrastructure behind your software without becoming dependent on expensive platforms, specialist teams, or increasingly complex cloud services?
Surviving AI with Minimal Viable Kubernetes is a pragmatic guide to running Kubernetes without the enterprise complexity that has traditionally surrounded it.
Written from more than 20 years of hands-on infrastructure and software experience, Jon Brookes explores how AI can help developers and small teams become more capable generalists, while using Kubernetes as a practical foundation for self-hosted and cloud-independent workloads.
At the heart of the book is Minimal Viable Kubernetes (MVK)—an opinionated approach that strips Kubernetes back to what a small business, startup, solo developer, or edge deployment actually needs.
The book focuses on the parts of Kubernetes that are often hidden behind managed services and simplified tutorials:
- Ingress — getting users reliably to your services
- Persistence — making application data survive beyond individual workloads
- etcd and cluster state — understanding what makes Kubernetes itself work
- Workload deployment — treating applications and infrastructure as replaceable rather than precious
- Backups and recovery — proving that your data can actually be restored
- Scaling — moving from a single-node deployment to a minimal multi-node cluster
- AI-assisted development — using AI as a force multiplier without surrendering your own knowledge and judgement
The book presents practical MVK configurations ranging from a minimal single-node deployment through to a three-node hyper-converged cluster, using technologies such as K3s, Kine, PostgreSQL, Traefik and Longhorn.
This isn't a book about building the biggest Kubernetes cluster possible.
It's about building enough.
It's for the founder who doesn't want a DevOps team before they have customers. For the developer who wants to understand what sits underneath managed Kubernetes. For the small business that wants predictable infrastructure costs and less vendor lock-in. And for the experienced infrastructure engineer who wants to cut through layers of unnecessary complexity.
The goal is simple:
Take back control of your infrastructure, without taking on more complexity than you need.
Minimal. Viable. Kubernetes.