llm-observability

Langfuse observability for DeepSeek Harness (dsh): exports agent sessions as OpenTelemetry trace trees (GenAI semconv) to Langfuse's OTLP endpoint

OpenTelemetry tracing for DeepSeek Harness (dsh): turns each agent turn into a GenAI span tree — steps, LLM calls with TTFT, tool executions, token usage — exported over standard OTLP to Jaeger, Grafana Tempo, SigNoz, Langfuse, or any compatible backend.

OpenTelemetry and Langfuse observability exporter for DeepSeek Harness: turn/step/tool/LLM spans, token and cost metrics, sanitized prompt/completion capture, async batching, bounded offline buffering, retry with backoff

X-ray for your DeepSeek Harness: context-tax attribution, entry inspection, dependency cascades

Persistent provider-reported token usage statistics for DeepSeek Harness with per-session, per-model, and daily activity data.

DeepSeek Harness usage dashboard with API balance, daily spend, external vision-call accounting, per-model stats, call logs, cache rate, TTFT, and CSV export.

Local request and response inspector for session-associated DeepSeek Harness LLM calls

DeepSeek Harness（DSH）只读执行轨迹复盘插件，支持规则分析、独立模型解读、证据定位、任务概览与运行对比。

Behavioural supervision for DeepSeek Harness agents: retry storms, reasoning-budget burn, and a backstop that survives host suspend.