Monitor and observe LLM applications in production. Covers token usage tracking and cost monitoring, latency monitoring (TTFT, TPS, E2E), LangSmith tracing, LangFuse integration, Phoenix/Arize for LLM observability, prompt/completion logging, conversation tracking, quality metrics over time, error rate monitoring, rate limit tracking, model comparison dashboards, feedback collection, A/B test analysis, hallucination rate monitoring, and LLM-specific alerting. Use when monitoring LLM applications, tracking costs, debugging quality issues, or setting up LLM observability infrastructure.
