Live workshop: instrument your AI agent with OpenTelemetry and debug it via Langfuse traces
Al_Grigor · x · 2026-11-03
AI Shipping Labs is running a hands-on tracing workshop that tackles a common pain point: when an agent returns a wrong answer, logs and the final output alone can't tell you whether the search, tool call, prompt, or model was at fault.
The workshop covers:
- Instrumenting the FAQ agent from earlier sessions with OpenTelemetry so every LLM call and tool call becomes a span in a single trace
- Sending traces to Langfuse (open source, self-hostable) and reading inputs, outputs, latency, tokens, and per-step cost
- Reusing the same OpenTelemetry traces with other backends
- Using a trace to pinpoint why one specific answer went wrong
- Turning interesting traces into eval examples
Participants need Python, an LLM API key, and a basic agent with tools; attendees leave with a working tracing setup for their own projects. Hosted by Alexey Grigorev, members-only.
More from coding & agent
- GenReasoning claims first public recipe for training agent collectives, hard-to-verify RL next — rosstaylor90 · 2026-11-03
- GitHub org/repo secrets can't be secured — use environments instead — AccBalanced · 2026-11-03
- Dev shows Codex refactoring his game duo, proposing typed equipment APIs and staged cleanup — Dimillian · 2026-11-03
- Inference CEOs say they're doing SEO for AI agents that buy compute, not just humans — abhijithneil · 2026-11-03
- Morpheus wins OpenAI London DevDay hackathon among 400+ teams by simulating your home for robots — Scobleizer · 2026-11-03
- Vibecoding discourse is hypnotizing normal people into paying for tools they don't need — arthurcolle · 2026-11-03