Reading 100+ Kubernetes Post-Mortems
summary
Review of 100+ postmortems finds a few concrete recurring failures: default CronJob settings spawn thousands of pods by missing concurrencyPolicy, mis-placed or malformed YAML leaves resources unregistered, missing CPU/memory limits lets a single third‑party pod OOM nodes, and a wildcard '' ingress* routes entire cluster traffic to one service. Root causes are absent validation and policies; the working mitigation is CI/CD–integrated policy enforcement (example: Datree) that blocks invalid Helm/Kubernetes YAML before it reaches the cluster.
AI-written summary · about
chapters (6)
related talks
33mWhat We Learned from Reading 100+ Kubernetes Post-Mortems
34mReading 100+ Kubernetes Post-Mortems - lessons learned
30mWhat we learned from reading 100+ Kubernetes Post-Mortems
36mKubernetes Confessions: Real Mistakes We’ve All Made (and What They Teach Us)
37m10 Ways to Shoot Yourself in the Foot with Kubernetes, #9 Will Surprise You
35m10 Weird Ways to Blow Up Your Kubernetes
35mKubernetes, Lessons Learnt
27mRunning Kubernetes in Production A Million Ways to Crash Your Cluster