Get hands-on building end-to-end observability and troubleshooting multi-agent AI systems
Agentic AI apps don’t fail silently; they fail distributedly. This hands-on lab shows how to make multi-agent AI systems observable end-to-end on Red Hat AI. By exploring a multi-agent app integrated with Model Context Protocol (MCP) tools, we demonstrate an AgentOps discipline that delivers visibility from infrastructure to AI model backends. As organizations scale, teams face cascading errors, hidden latency, and complex integration bottlenecks. Attendees will leave with practical patterns for observing, troubleshooting, and evaluating agentic workflows.
Attendees will also learn how to:
- Monitor the stack: Use Red Hat AI’s out-of-the-box observability stack to track key metrics and logs.
- Trace multi-agent executions: Track requests across multi-agent frameworks (like LangGraph/LangChain) and MCP tools to understand the complete decision-making path using Red Hat AI and MLFlow tracing features.
- Diagnose and fix distributed failures: Simulate real-world rollout issues, such as artificial delays in MCP servers or failing multi-agent communications, and use tracing to pinpoint the root cause and deploy the fix.
- Large language model (LLM) evaluations: Go beyond basic observability by combining tracing with LLM evaluations in MLflow, ensuring your agents maintain high-quality outputs alongside system reliability.
No AI experience is required.
Who should attend:
- Data engineers
- Data scientists
- ML engineer
- Infrastructure architects
- Infrastructure specialists
- Enterprise architects
- Developers
- Application architects
- Developer team leads
Know before you go: Operations, developers, and data science practitioners taking part in these hands-on lab should have these suggested tools and knowledge of these areas.
- A laptop computer running Windows, MacOS, or Linux with the Firefox or Chrome web browser.
- Entry-level Kubernetes concepts
- A general understanding of Linux containers (e.g., Docker, CRI-O, etc.).
- General knowledge of AI terminology (e.g., machine learning, deep learning, foundation models, etc.).
- No AI experience is required.
Virtual event details
Date: September 23, 2026
Time: 1:00 PM - 3:30 PM ET
Any questions? Please email erhernan@redhat.com
