TtTRz/dsh-langfuse

README 摘要

dsh-langfuse Langfuse LLM observability for DeepSeek Harness — one OpenTelemetry trace tree per agent session, with feedback scores and subagent lineage. Send every agent session to Langfuse as an OpenTelemetry trace tree — turn → trace , model step → generation , tool call → tool span — with GenAI semantic-convention attributes, token usage, time-to-first-token, and error statuses. ✨ Features - 🧵 Full session tracing — every turn, generation, and tool call lands in Langfuse with model, provider, usage (incl. cache-read & reasoning tokens), and TTFT. - 🚚 Content-Length transport — spans are delivered as a single write with an explicit Content-Length header, never chunked. Some Langfuse deployments sit behind gateways that mangle chunked POST bodies ( 400 Failed to parse OTel JSON Trace ); this transport sidesteps that class of bugs entirely. - 👍 Feedback scores — /feedback records become TEXT scores on the session's latest turn trace. - 🔗 Subagent lineage — child-session turn traces link to the parent session's trace (same-process). - 🎚️ Three sharing modes — FULL (live), FEEDBACK ONLY (export only when the user records feedback), DISABLED (default; nothing leaves the process)…

分类