kubernetes on AWS gone wild
summary
Nine war stories about running kubernetes on AWS cover concrete fixes not platitudes: kube‑up breaks on updates, switch to kops, save a configured node as an AMI for fast recovery, use IAM server roles and per‑account isolation to avoid S3 owner and permission nightmares, expect network bandwidth to become the bottleneck as microservices scale, span masters across availability zones for redundancy, prefer managed DBs and external Elasticsearch, and patch a custom registry HTTP/2 timeout (900s) to stop unexplained pull failures.
AI-written summary · about
chapters (9)
- 0:00Start
- 1:16THREE MAIN TOPICS
- 1:57AUTOMATED VS. MANUAL
- 13:44PERMISSIONS ON AWS Let's talk about compliance
- 20:52KUBE-UP & COPS
- 23:04NETWORKING
- 25:42DATABASES
- 27:13LOGGING know what is going on
- 29:17CUSTOM REGISTRIES Not everything is supposed to be public
related talks
35mKubernetes on AWS: Practices & Opinions
27mRunning Kubernetes in Production A Million Ways to Crash Your Cluster
33mKubernetes Security: Horror Stories to Help Secure Your Cluster
42mMastering Kubernetes: Challenges, Strategies, and Best Practices
29mRunning Kubernetes on 'small scale' clusters - lessons learned
35mKubernetes on AWS: Zero to hero
17mKeynote: Kubernetes: Lessons from the Trenches
22mThe House is on fire: mysteries of on-call K8s engineering