Observability and Agentic AI - From Observability to Answers and New Challenges

Explore the evolution of observability from manual data querying to AI-powered automated insights in this 32-minute sponsored conference session. Trace the journey from traditional log searching and dashboard building to modern structured telemetry systems, examining how the field has remained largely manual, fragmented, and reactive despite advances like OpenTelemetry and query languages such as SPL and PromQL. Discover the key technical requirements for next-generation observability including wide events, unsampled high-cardinality data, and fast scalable query engines. Learn how agentic AI systems can transform observability workflows by guiding investigations, interpreting anomalies, and enabling natural language troubleshooting grounded in real telemetry data through standardized model and prompt benchmarks against sample datasets. Understand both the transformative potential of AI in observability and its current limitations, including critical challenges associated with relying on autonomous agents for system monitoring and incident response.