15 KUBERNETES FAILURE POINTS YOU SHOULD WATCH
summary
Claims a concrete checklist: 15 failure points across applications, nodes, scheduler, deployments, pods, control plane, DNS and human error. Demonstrates metrics to collect (latency traffic errors saturation), example alerts and dashboards built with Prometheus + kube-state-metrics, and uses event counts and pod restart metrics to detect crashes. Highlights gaps — persistent volume visibility is weak — and shows a troubleshooting workflow using an inspect syscall capture to diagnose CrashLoopBackOffs.
AI-written summary · about
chapters (33)
- 0:00Intro
- 1:38Monitoring & Security Platform for Containers
- 1:59Monitoring 15 Kubernetes failure points
- 3:25USE method
- 5:05The holy service metrics
- 6:30Kubernetes orchestration
- 6:49Kubernetes hierarchy
- 7:44Kubernetes metadata: labels
- 8:14Leverage metadata (by service)
- 8:28Leverage metadata (by pod)
- 10:03Container scheduling
- 11:16node availability
- 13:35Container resource requirements
- 14:24CPU resources 6. memory resources
- 15:38Container disk requirements
- 16:13disk resources
- 17:47Container orchestration
- 19:12running instances
- 19:29desired instances
- 20:55deployment updates glitches
- 22:39Container livecycle state
- 24:15Liveness probes
- 25:42Ready-ness probes
- 27:18pod status
- 28:59pod restarts
- 30:09CrashLoopBackoff event
- 30:20Sysdig Inspect
- 30:49Kubernetes internals
- 32:37APIserver
- 33:24KubeDNS / Istio
- 34:39What are we deploying?
- 35:19monitor your commands
- 38:02Recap (2)
related talks
05mLightning Talk: The Top 5 Kubernetes Metrics to Monitor
36mKubernetes Confessions: Real Mistakes We’ve All Made (and What They Teach Us)
20m10 Common Things People Do Wrong in K8s Environments
35mKubernetes, Lessons Learnt
35m10 Weird Ways to Blow Up Your Kubernetes
36m101 Ways to Crash Your Cluster
29mWhen Kubernetes Misbehaves – Pitfalls and Lessons Learned
37mWeighing a Cloud: Measuring Your Kubernetes Clusters