A Kubernetes cluster is not a platform. It is a starting point. Most teams I work with have AKS or EKS up and running within a sprint, and then spend the next two years figuring out everything that sits on top: ingress, certificates, secrets, observability, backups, disaster recovery, and the developer-facing pieces that let people ship code without filing tickets.

This session is a live walkthrough of what those layers look like in practice. We will start from a clean cluster and build up the parts that turn it into something a developer can use safely, with the operational guardrails an IT team can actually defend. I will break a few things on purpose along the way, because the failure modes are usually more instructive than the happy path.

Most of what I will show comes from running NForza Core, the Kubernetes platform we built for our own customers and have been operating for about ten years. I will be honest about the choices that worked, the ones we changed our minds about, and the ones that still cost us sleep.

If you are running clusters and the gap between "it works" and "developers love using it" feels bigger than it should, this session is for you.