Deep Dive To AI Agent Observability - Guangya Liu, IBM & Karthik Kalyanaraman, Langtrace AI
Guangya Liu, IBM, Karthik Kalyanaraman, Langtrace AI
KubeCon + CloudNativeCon Europe 2025 · Session
This talk, presented by Karthik Kalyanaraman from Langtrace AI (with contributions from Guangya Liu of IBM), provides a comprehensive exploration of **AI agent observability**. It addresses the significant challenges arising from the rapid evolution of AI architectures, particularly with the emergence of large language models (LLMs), vector databases, and sophisticated orchestration frameworks. The core premise is that traditional software observability paradigms are insufficient for the non-deterministic, complex, and rapidly changing landscape of generative AI applications and multi-agent systems.
AI review
This talk provides a critical and timely deep dive into AI agent observability, outlining the paradigm shift required for monitoring the non-deterministic, complex, and rapidly evolving multi-agent systems. It meticulously details the OpenTelemetry GenAI Semantic Conventions as the foundational standard, explaining unique signals, tracing methodologies across LLM, storage, and framework layers, and the profound implications for reliability, cost, security, and continuous improvement. The speakers, directly involved in shaping these standards, offer invaluable, actionable insights for moving…