Agents in Kubernetes 101: A Crash Course on Cloud Native AI Agents
summary
Explains how a Kubernetes-native agent stack actually works and what was built: a declarative Kagent runtime that mounts LLMs, tool plugins and agent-to-agent plumbing; a live demo fixes a “route not found” by having an agent call Kubernetes tools and apply a corrected HTTPRoute manifest; nested agents include a Helm agent, a Kubernetes agent and an MCP server that lists real GitHub PRs; protocols covered are MCP and A2A, and the setup runs against a local Llama model and production-scale clusters (thousands) for ops automation.
AI-written summary · about
chapters (18)
- 0:00AI agents in Kubernetes intro
- 1:06Agents need tools and memory
- 2:30Core agent components
- 4:15Tool integration challenges
- 5:21Memory and orchestration
- 6:01Kgent framework intro
- 6:45Live demo setup
- 9:12Debugging with Kgent
- 11:59Interacting with agent
- 14:04Engagement demo
- 15:28Multiple specialized agents
- 16:11A2A protocol
- 17:18Nested agents in Kubernetes
- 18:42Testing multi-agent workflows
- 20:49Kgent features
- 21:51CNCF agent projects
- 22:11Resources and community
- 22:52Q&A
related talks
27mBeyond Chat: Bringing Agents To Kubernetes
06mFrom Containers To Agents: The Next Cloud Native
30mSo You Want To Run AI Agents on Kubernetes: A 101 Guide
40mAgents in Action: Advancing Open Source AI Missions
14mKeynote: Key Cloud Native Technologies in its Next Decade
27mFrom Chatbots to Agents: Orchestrating Agentic AI on Existing Enterprise Kubernetes Infrastructure.
32mBuild your own Autonomous Agent powered by LLMs using Kubernetes
25mWhat We Learned Securing AI Agents at Scale on Multi-Tenant Kubernetes Clusters