In this video, we run an AI agent inside a live Kubernetes cluster with kagent and use it to debug two real failures that leave nothing in the logs.
? Try the free hands-on lab here: https://kode.wiki/4eZV5e1
We cover every piece: how Kubernetes services find their pods through labels, the selector mismatch that silently drops all your traffic with zero errors, the one kubectl get endpoints command that exposes it in seconds, what kagent actually is and how it runs as a Kubernetes resource you define in YAML, the scoped tools you allow it to touch, and a live before-and-after where we fix a broken service and a stuck autoscaler by hand first, then hand the exact same problems to the agent and watch it connect every dot on its own.
⏰ Timestamps:
00:00 - Introduction: healthy pods, no traffic
00:35 - How Kubernetes services find pods with labels
01:32 - What kagent is and how it runs in your cluster
02:38 - The lab: before and after
03:01 - Fixing the order-api service by hand
04:22 - Fixing the inventory HPA and resource quota by hand
06:11 - kagent debugs the service in one prompt
07:24 - kagent traces the autoscaler failure
08:07 - Quiz - By hand vs kagent
? Build it yourself with our free lab: https://kode.wiki/4eZV5e1
? Go deeper with the full kagent course: https://kode.wiki/4pdO9yI
? Like if this clicked, and subscribe for more hands-on Kubernetes, DevOps, and AI engineering.
#kagent #Kubernetes #AIAgents #k8s #DevOps #AgenticAI #CloudNative #kubectl #KubernetesTutorial #AIforDevOps #Troubleshooting #CNCF #SRE #PlatformEngineering #MCP #KubernetesTroubleshooting #KodeKloud