Demystifying DNS: A Guide to Understanding and Debugging Request Flows in Kubernetes Clusters

on YouTube ↗32m

" summary

Shows how to trace DNS from pod → node → CoreDNS → upstream using real commands and logs: enables CoreDNS log plugin to record queries, runs Hubble (with Cilium L7 proxy) to see flows, and uses Inspector Gadget DNS gadget (eBPF) to capture OS+Kubernetes context. Demonstrations resolve example.com and unknown.example.com, reveal success and NXDOMAIN, reproduce a timeout by inserting CoreDNS eratic drop, and confirm the failure occurs inside CoreDNS; notes no DNS-over-TCP support yet.

AI-written summary · about

" chapters (9)

1. 0:00Introduction to DNS debugging
2. 1:27Understanding DNS components
3. 3:36Using CoreDNS log plugin
4. 4:56Observability with Hubble
5. 8:43Inspektor Gadget framework
6. 16:43Scenario: Upstream server health
7. 20:39Scenario: Identifying failures
8. 24:11Conclusion and summary
9. 25:18Q&A session

" related talks

• 38mIntro + Deep Dive: CoreDNSYong Tang · KubeCon + CloudNativeCon China 2019
• 06mLightning Talk: Is Your Kubernetes Cluster's DNS Working?Jonathan Perry · KubeCon + CloudNativeCon North America 2019
• 38mCoreDNS: Intro and Deep DiveJohn Belamaric, Yong Tang · KubeCon + CloudNativeCon Europe 2022
• 41mService Discovery With Hybrid and Multi-Cloud: Introduction to CoreDNSYong Tang · KubeCon + CloudNativeCon North America 2019
• 29mIntro: CoreDNSDaniel Garcia, Michael Grosser · KubeCon + CloudNativeCon Europe 2019
• 34mIt’s Always DNS: So Let’s Break It On PurposeJoel Studler, Georgios Daskalopoulos · ContainerDays Hamburg 2026
• 38mIntro + Deep Dive: CoreDNSYong Tang · KubeCon + CloudNativeCon + Open Source Summit China 2019
• 35mDeep Dive: CoreDNSJohn Belamaric, Yong Tang · KubeCon + CloudNativeCon Europe 2019