LLM monitoring in Sentry tracks every LLM call, agent run, and tool call in production with errors, latency, and token cost, tied to the code that triggered it. Each call is captured as a trace with full prompt and response context, so you can debug failures and optimize performance without leaving your application monitoring. See how to monitor AI agent costs in Next.js.
Track all agent runs, error rates, LLM calls, tokens used, and tool executions. Monitor traffic patterns and duration metrics across your AI-powered features.