Building a Deterministic vCluster Validation MCP Server to Ground AI in Real Schemas

You ask an AI to generate a Kubernetes manifest, Helm chart values, or Ansible playbook. It responds instantly with clean, well-formatted YAML. You apply it. Nothing works.

This isn't a bug—it's AI hallucination. The AI knows YAML syntax but hallucinates config options that don't exist, mixes incompatible versions, or confidently suggests deprecated fields. It generates what looks right based on patterns, not what is right according to actual schemas.

Photo by Growtika on Unsplash

In today's cloud-native landscape, GPU workloads are becoming increasingly critical. From training large language models to running inference APIs, organizations are investing heavily in GPU infrastructure. But with this investment comes a challenge: how do you safely test and deploy new GPU schedulers without risking your entire production environment?

Related talks: Watch my SREDay Paris Q4 2025 talk on this topic. Also presenting at Conf42 Kube Native 2025. Check the talks page for more details.

How to learn and stay up to date with DevOps and Cloud Native technologies​

Tell me and I forget. Teach me and I remember. Involve me and I learn.

Bejnamin Franklin

Know what is important to know​

A famous Greek philosopher Aristotle (384 BC — 322 BC) has been called the last person to know everything there was to know in the science. Since than science grown exponentially, became divided and subdivided into specialized, narrow disciplines.

I have given a talk about this topic at the WorldWide Architecture Summit 2 on the 3rd of August 2021. You can find the talk recording here.

Photo by Christina @ wocintechchat.com on Unsplash

Static websites have become a mainstream content delivery medium consumed by modern browsers. In this blog, we will look at different ways of deploying a sample static website using automated CI/CD pipelines and different hosting providers.