AI/ML
///Advanced
AI Observability & Tracing
Instruments LLM applications for observability: traces and spans across model, tool and retrieval calls, token and cost tracking, latency breakdown, quality metrics and user feedback, alerting on drift and failures, and dashboards for debugging. Use when the user runs an LLM app in production, wants to trace a bad response, track spend and latency, or set up tracing and evals across a stack.
Preview
# AI Observability & Tracing LLM apps fail in ways classic services don't: a bad response can come from a prompt, a model version, a retrieval step, or a tool call. This skill makes those paths visible by instrumenting every step into traces, then wiring metr
Related skills