k8snetlook – Root-Causing k8s Network Problems in an Automated Way
Talkat KubeCon + CloudNativeCon North America 2021 · on Wednesday, 2021-10-13· at 14:30· Room 403 AB + Online·
summary
k8snetlook automates the usual manual checklist by running on the node that hosts the source pod, entering the pod network namespace to pod-side imitation, and executing host checks (gateway, API servers), DNS lookups, per-endpoint probes and path MTU discovery. demoed failures include a network policy blocking east‑west traffic and MTU mismatches fixed by lowering MTU to 1480; tool reports partial degradation (e.g., 18 of 20 endpoints reachable), runs as a binary or Docker image, and successfully pinpoints the failing layer.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 6:20Communication across the cluster
- 8:51Network issues
- 11:37Big hammer approach
- 14:00Troubleshooting workflow
- 20:32Open source tools
- 23:15What to use
- 23:54Automated way
- 24:05Why k8snetlook
- 25:48What can it do
- 26:46Demo
- 33:20Outro
related talks
34mIt's Not Always the Network (But Here's How to Prove It): Kubernetes Packet...
33mThree Surprising K8s Networking “Features” and How to Defend Against Them
27mDeconstructing K8s Networking for Knowledge and Profit
32mUnderstand and Troubleshoot the “Magic” of Kubernetes Networking - Minhan Xia & Rohit Ramkumar
28mHow to Break your Kubernetes Cluster with Networking
39m15 KUBERNETES FAILURE POINTS YOU SHOULD WATCH
30mStreamlined Troubleshooting in Kubernetes
40mThe Network in Kubernetes: Can We See It More Clearly?