GenAI Framework Observability

Explore generative AI framework observability through practical code examples in this 31-minute conference talk from Cloud Native Summit 2025. Learn how GenAI frameworks enable users to interact with applications using natural language, images, or sound instead of traditional forms, while providing flexibility in Large Language Models (LLMs) and orchestration of tasks like RAG or tool calls. Discover the critical need for deep observability due to the unpredictable nature of LLMs. Examine hands-on code demonstrations using Spring AI (Java), Vercel AI (Node.js), and Semantic Kernel .NET frameworks. Master the implementation of Elastic Distributions of OpenTelemetry (EDOT) to capture comprehensive logs, metrics, and traces without modifying your core application code. Gain actionable techniques for iterating on and understanding your GenAI applications in production environments.