Fixing Potholes And Broken Roads: Devs Can Troubleshoot the Network With Ops Too!
summary
Local container fails because the app binds to localhost instead of the host network so changing localhost to 0.0.0.0 fixes the curl connection; Kubernetes ingress reports “no healthy upstream” because selectors mislabel pods so fix Kubernetes selector label restores endpoints; requests time out under load when a pod is capped at a quarter CPU and 0.25 GB RAM so adding an allow rule to NetworkPolicy and increase CPU and replicas or enable an HPA removes the bottleneck.
AI-written summary · about
chapters (12)
- 0:00Welcome
- 1:53Devs troubleshoot networks
- 4:23Common network errors
- 6:28Local dev tools
- 8:58Platform tools limits
- 11:46Enterprise traffic tools
- 12:31Fix Docker network
- 14:16K8s service debugging
- 16:22Resolving Kubernetes network policy blocks
- 17:24K8s workload handling
- 20:34Key takeaways
- 22:00Closing advice
related talks
34mIt's Not Always the Network (But Here's How to Prove It): Kubernetes Packet...
32mUnderstand and Troubleshoot the “Magic” of Kubernetes Networking - Minhan Xia & Rohit Ramkumar
42mKubernetes Networking at Scale
35mKubernetes, Lessons Learnt
38mLogs Told Us It Was DNS, It Looked like DNS, It Had to Be DNS, It Wasn't DNS
32mCruise’s Self-Driving Networking Journey
29m“Are These Things Talking to Each Other?” - Observing Kubernetes Networking...
34mReading 100+ Kubernetes Post-Mortems - lessons learned