Building a Fault-Tolerant Resources Controller on Kubernetes
summary
Custom Kubernetes controllers are implemented as CRDs that create deployments and services at runtime; a working demo runs on a three‑node cluster with two controller replicas where one holds a lock and the other is passive, and failover after killing the leader takes seconds rather than the default ~5 minutes. Uses leader election with client-go via a ConfigMap/Endpoint lock, demonstrates a controller vs control‑II pattern, warns about missing CRD validation in older versions and starting watches only after winning the lock.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 0:54Overview
- 2:00History
- 5:08Leader Election
- 7:23How it works
- 9:00Demo
- 21:18Controller vs Control II
- 25:06Demo II
related talks
29mKubernetes Custom Resources with Kubeless and Metacontroller
40mExtending K8s- Introducing Custom Resources, Controllers, Operators & Schedulers
27mDemystifying Why the World Is Built on Kubernetes: Learning To Leverage Bespoke CRDs and Controllers
26mExtending Kubernetes with Custom Controllers
34mPractices for Designing Kubernetes Controllers and CRDs — A Networking Example
31mBeyond the Limits: Scaling Kubernetes Controllers Horizontally
48mKubernetes as backend: CRD & Operators
46mOperators Live in the Past: Designing Reliable Kubernetes Controllers