LLM Observability

Parent: AI Agent Ecosystems · researched 2026-05-31T17:21:52.875Z· 31 sources · 12 concepts · skill ai-agent-engineering

AI & agent-engineering family ROUTER. Split into: ai-agents-orchestration (agent frameworks, multi-agent, memory, planning, guardrails, coding/GUI agents, autonomous loops, eval); ai-rag-retrieval (RA

ai-agent-engineering

Children

Frontier under this node: Agent trajectory and tool-call observability, Answer-relevance monitoring, Datasets and experiment tracking, Hallucination / groundedness / faithfulness monitoring, LLM-as-judge production scoring, LLM/agent tracing and spans, Loop / recursion detection in agents, Observability tooling landscape (Langfuse, LangSmith, Arize Phoenix, WhyLabs/LangKit, Helicone, Datadog, OpenLLMetry), Online (reference-free) evaluation vs offline eval, OpenInference span-kind standard, OpenLLMetry / Traceloop instrumentation, OpenTelemetry GenAI semantic conventions, PII detection and redaction at runtime, Prompt / output / embedding drift detection, Prompt management and versioning, RAG observability (RAGAS: context precision/recall, faithfulness), Runtime guardrails and safety monitoring, Token, cost, and latency monitoring

← the whole tree · 3D view· how to read this page