In this session we will look at how proper observability and instrumentation is becoming ever more important in the new emerging agentic architectures.

You will learn how to include telemetry data from agents in the overall signals from your application landscape. We are going to dive into architectural patterns to help collect data from both the agentic and non-agentic parts of your system. Also, you will see how to enrich the telemetry data from your agents and language model conversations to be OpenTelemetry compliant. Finally, we will have practical advice on the translation of raw signals from AI elements to meaningful understanding of the performance and health in the overall architecture.

After this session you will have deeper insights in the way observability and instrumentation is used with the new AI elements inside your system and get full visibility in its performance, health and business value.

You will learn:

• To understand and implement instrumentation in agentic systems for observability
• How to design modern observable architectures and topologies with agents and language models
• How to translate instrumentation data and signals to meaningful insights on agent performance and health