Kubernetes Isn't Outdated. It's Evolving. And So Should You.
Kubernetes is not out of trend; it is infrastructure. What is changing is the role engineers need to play around it, from cognitive load to platforms, AI and MCP.
$ cat ~/intro.txt
I build and modernize cloud-native platforms: Kubernetes, AWS, GitOps, observability and the developer experience on top. Ten-plus years of GNU/Linux, a bias for automation, and the belief that res publica non dominetur.
series · 27 posts
A daily series from Kubernetes basics to building Internal Developer Platforms.
series · 8 posts
Factory, Facade, Observer, Command and friends, mapped to the decisions you make building platforms.
series · 7 posts
Lo que Platform Engineering necesita entender para habilitar pipelines de Machine Learning.
series · 4 posts
Estrategia, criterio y ejecución para modernizar sin romper la operación.
Kubernetes is not out of trend; it is infrastructure. What is changing is the role engineers need to play around it, from cognitive load to platforms, AI and MCP.
The closing post of the series. The thread that connects all six patterns, the evolution from glue spaghetti to platform control plane, and why having architectural vocabulary changes how platform teams make decisions.
Hardcoding deployment strategy into pipelines creates a hidden coupling that makes platforms less reliable. The Strategy pattern explains how to make canary, blue/green, and progressive delivery first-class platform concerns.
Every platform connects systems that weren't designed to talk to each other. That connective tissue is glue — and in most platforms, it's 60-80% of what the team actually builds. Here's what glue is, where it lives, and how to design it intentionally.
Pipelines that run kubectl apply directly are automated but invisible. The Command pattern explains why GitOps is architecturally sound — not just a convention — and what you gain from separating the definition of a change from its execution.