Stop starting to use Kubernetes (or why Kubernetes is just the worst)

Build & Deploy

Kubernetes is just the worst. Pick something that will give you more time, not take away your time. Don’t start to use Kubernetes now. 

And if you’re already using Kubernetes, are you really on top of all that BAU maintenance work? Or as I prefer to call it, ‘yaml at scale’. 

Why is it a bad idea to start with Kubernetes? How can you enable your team(s) to focus on higher-value work, rather than BAU work? How can you deliver more sooner? Why does Kubernetes get in the way?

Choosing a compute model is hard. They can have a long lifecycle, and migrating to a new model can force you to re-architect your world. You want to pick a compute model that serves your needs whilst minimising maintenance and BAU work.

What options are available so you can have your cake and eat it? Which tool is right for the job? Why should you start with functions and move to lower abstraction levels when absolutely required? Why is Kubernetes the worst?

I’ve implemented Kubernetes at scale in three organisations and lower complexity alternatives in two others. I’ll cover why starting with Kubernetes in 2022 is a bad idea, why picking Functions as a Service or Containers as a Service tools will help you get to higher-value work faster, and how to identify which compute model is best for you.

Adam Hansrod

Equal Experts

Adam is a consultant with Equal Experts, and focuses his time helping organisations solve scaling challenges in a sustainable way.  In practice it means he spends half his time talking to people and half of his time typing to build services.

Adam has spent time building high-performance financial trading systems, iterating on digital tooling and processes to support hordes of teams,  and helping organisations build an engineering culture from scratch. He's been fortunate enough to have led award-winning platform teams that help other teams deliver more in less time for a lower total cost of ownership.