End-to-end monitoring for LLM applications — trace prompts, evaluate responses, track costs, and debug AI pipelines in…