Kubernetes Day 2: Cluster Operations [I]
summary
Shows a working workflow for long‑term clusters: uses Tectonic installer on AWS, runs Kubernetes on Kubernetes as self-hosted Kubernetes, manages the critical datastore with a custom etcd operator that auto‑scales a 5‑member etcd quorum, restores a terminated master by rebuilding a member from peers, and uses the Prometheus/ kube-prometheus stack for alerts on etcd quorum, API latency and scheduling delays.
AI-written summary · about
chapters (12)
- 0:00Start
- 1:51Discussion Outline
- 19:00API Server: Surviving Failure
- 26:09Scheduler and Controller Manager: Falure Risks
- 27:01Kubernetes DNS
- 28:35Node Selector: Pin Pods to Label Query
- 29:00Failure Domains (beta)
- 30:41Prometheus Operator
- 31:31Node Status: Alerts
- 33:13Backup and Disaster Recovery
- 33:47DEMO: Upgrade/Downgrade Scheduler
- 34:18Scaling Nodes
related talks
06mKeynote: Distributed Systems Simplified Using Kubernetes
30mManaging Kubernetes: What You Need to Know About Day 2 - Craig Tracey, Heptio
47mKubernetes Day 2 to 5
35mKubernetes journey
45mKubernetes Operability Tooling
03h 41mKubernetes Basics - a live tutorial
36mKubernetes with a single master
29m11 5 years of kubernetes and looking to the future