Unlocking Kubernetes Intelligence: Deploying AI Foundation Mode...
summary
K8sGPT runs as a CLI or Kubernetes operator (installed via Helm) and scans clusters, then sends selected errors to an AI backend; the demo uses K8sGPT CLI with a SageMaker backend (deployed with CDK using Llama 2) and discovers six problems in one run — node health, bind to Port 18 blocked by dropped capability, insufficient CPU request (250 → lowered), and a file path typo in Nginx; deploying the endpoint takes ~460s via CDK deploy, fixes work, and the backend can stay inside the network for privacy.
AI-written summary · about
chapters (12)
- 0:00Kubernetes intelligence with AI
- 1:01Large language models basics
- 3:07AI for technical professionals
- 4:54AI automation and privacy
- 6:42Live demo: K GPT for Kubernetes
- 19:23K GPT features and contributions
- 20:05Installing K GPT locally
- 22:38Deploying K GPT in production
- 24:03AI backends in K GPT
- 27:48Deploying AI backend with AWS CDK
- 32:44Key takeaways
- 34:09Q&A
related talks
06mLightning Talk: AI-Powered Kubernetes Diagnostics With K8sGPT
34mSolving Kubernetes Mysteries with K8sGPT: Your AI-Powered Troubleshooting Companion
29mGain SRE Superpowers with K8sGPT
09mLightning Talk: Sailing Kubernetes Operation with AI Power
09mTransform Your Kubernetes Cluster Into a GenAI Platform: Get Ready-t...
38mAI in Kubernetes: How to Get Started?
09mKeynote: Kubernetes for GenAI: The New Operating System of the AI-First World
12mUnleashing Kubernetes Intelligence - running k8sgpt utilizing your own fine-tuned LLM