Kubernetes isn't the answer
summary
Argues Kubernetes often causes more organizational friction than it solves and recommends choosing the simplest tool that fixes a specific bottleneck. Gives concrete alternatives: use Nomad for simple on‑prem job scheduling, consider serverless or a managed Kubernetes offering instead of DIY, and limit Kubernetes features (start with a single cluster and deploy only what you need). Shows how to prove org problems with value stream mapping to expose handovers and wait time and targets adoption steps, not big‑bang rewrites.
AI-written summary · about
chapters (40)
- 0:00Intro
- 1:13Who is Joe
- 2:00Lets begin
- 2:30Why we are here
- 2:56Kubernetes isnt always the answer
- 3:25Technology decisions the wrong way
- 4:07Kubernetes has lots of potential
- 5:12Potential is unrealized success
- 6:16Kubernetes analogy
- 7:52Why are we doing Kubernetes
- 9:12Reducing complexity
- 10:06The right choice
- 11:29The bottleneck
- 11:51What do we know
- 12:35Minimalism
- 13:10Kubernetes diagram
- 14:08The organizational piece
- 14:42Dont do just technology
- 15:13Solve organizational issues
- 15:51Examples
- 17:01Outsourcing
- 18:40Ticketing systems
- 19:46Handovers
- 21:01Dependencies
- 22:03Responsibility
- 22:51Stakeholders
- 23:30Onprem
- 24:14DIY
- 25:38Lead time
- 26:54Work in progress
- 27:54The solution
- 28:40Vending machine model
- 31:41Value stream mapping
- 34:19The goal of this approach
- 35:24Kubernetes will solve everything
- 36:01Improve rework or prevent it
- 37:08Technical debt
- 39:36Boy scout rule
- 41:57The Hero Complex
- 48:31Can Kubernetes solve organizational issues
