From Logs To Insights: Real-time Conversational Troubleshooting for Kubernetes With GenAI
Talkat Observability Day 2025 · on Tuesday, 2025-04-01· at 11:20· Level 1 | Hall Entrance S10 | Room C·
summary
Shows an open-source RAG pipeline that ingests systemd logs, Kubernetes events and app logs via Fluent Bit into Amazon Kinesis, triggers Lambda to encode with Amazon Titan embeddings and store vectors paired with text in Amazon OpenSearch. A chatbot generates parsed kubectl commands, executes read‑only kubectl against the cluster, re-queries the model with outputs and returns root‑cause steps; demo fixes a pod pending on a non‑existent ConfigMap and an image pull issue. Code deploys with Terraform.
AI-written summary · about
chapters (11)
- 0:00Real-time troubleshooting intro
- 1:02Chatbot demo
- 4:32Problem and solution
- 6:17Kubernetes observability challenges
- 7:42Solution architecture
- 9:48Customizing foundational models
- 12:57How RAG powers chatbot
- 14:41Agentic workflow enhancements
- 16:25Live demo walkthrough
- 20:15Code walkthrough
- 24:53Key takeaways
related talks
42mEffortlessly Build, Run, Secure, and Manage Traffic for a Generative AI Application From...
07mLightning Talk: Demystifying Kubernetes Observability with Generative AI and LLMs
13mExploring AI-Powered Chatbots and Their Transformative Role
42mAgentic AI with AWS Streaming
26mKubernetes and Logging: Do It Right
36mGENAI-Powered chatbot to increase team efficiency
01h 36mTutorial: From Planni...
09mTransform Your Kubernetes Cluster Into a GenAI Platform: Get Ready-t...