Deep Dive into Kubernetes Internals for Builders and Operators
summary
Builds a one‑node cluster from binaries: etcd, kube-apiserver, controller-manager, scheduler are started, then a deployment (nginx, three replicas) creates a ReplicaSet and pods; kubelet, docker, kube-proxy, CNI (kubenet then kube-router) are used to run pods, create iptables rules for ClusterIP services, and later add a second node. Networking breaks (duplicate pod IPs, missing routes, iptables blocking) are fixed manually; security and TLS for service accounts remain unaddressed.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 4:36Kubernetes Anatomy
- 10:30What do we need
- 11:23Live demo
- 12:05Starting the API server
- 23:42Adding more nodes
- 23:52Connecting to other nodes
- 26:18Scaling up
- 28:02Networking
- 30:52Pods
- 33:55CNI
- 36:00How to use CNI
- 38:05Demonstate
- 40:01Whats missing
related talks
44mDissecting Kubernetes (K8s) - An Intro to Main Components
44mDemystifying Kubernetes internals
27mKubernetes the Hard Way: A Live Journey
01h 27mKubernetes Networking 101
02h 42mWorkshop: Kubernetes 101
34mIntroduction to Kubernetes for Developers
31mBuilding Native Kubernetes Integrations with Operators
48mKubernetes Component by Component
