10 Common Things People Do Wrong in K8s Environments
summary
Lists 10 concrete Kubernetes mistakes and fixes: Pod Disruption Budget drain-lock causes stuck one-replica deployments during node drains; privileged containers allow host takeover via chroot and host mounts; missing resource requests and limits produce OOMs, noisy neighbors and broken autoscaling—use VPA and percent-based PDBs; absent liveness/readiness checks hide dead pods and cost six hours to diagnose; network policy defaults to allow-all; enforce images with Kyverno / Open Policy Agent and mirror registries in-cluster (spiegel) to avoid 200 pods stuck and $50k crypto-miner incidents.
AI-written summary · about
chapters (9)
- 0:00Introduction and Speaker Background
- 0:13Common Pitfalls in Kubernetes
- 1:30Pod Disruption Budget Issues
- 2:57Privileged Containers and Security Risks
- 4:44Network Policy and Resource Management
- 7:37Monitoring and Health Checks
- 9:37Scheduling and Image Trust
- 14:15Misuse of Pods as VMs
- 18:28Conclusion and Final Thoughts
related talks
36mKubernetes Confessions: Real Mistakes We’ve All Made (and What They Teach Us)
51mIntroducing the OWASP Top 10 for Kubernetes
35m10 Weird Ways to Blow Up Your Kubernetes
37m10 Ways to Shoot Yourself in the Foot with Kubernetes, #9 Will Surprise You
34mReading 100+ Kubernetes Post-Mortems - lessons learned
35mKubernetes, Lessons Learnt
39m15 KUBERNETES FAILURE POINTS YOU SHOULD WATCH
36m10 More Weird Ways to Blow Up Your Kubernetes