Datadog LLM Observability
Monitor, troubleshoot, and evaluate LLM apps inside the Datadog platform.
About
Datadog LLM Observability, now also branded Agent Observability in Datadog's own documentation, brings LLM and AI agent monitoring into the platform enterprises already use for APM, infrastructure, logs, and RUM. It traces LLM calls, fixed workflows, and agent steps, compares prompts and models in experiments, tracks token cost, and uses a Sensitive Data Scanner to redact data and flag prompt-injection attempts.
Expanded in June 2025 with AI Agent Monitoring, LLM Experiments, and an AI Agents Console, it supports the OpenTelemetry GenAI semantic conventions and auto-instruments frameworks like LangChain and OpenAI. Pricing meters LLM spans: a free plan covers 40,000 spans per month, with Pro starting around $160 per month for 100,000 spans. It is SaaS-only and fits teams that already run Datadog and want LLM traces in the same governance and billing framework.
What it can do
- Trace LLM calls, fixed workflows, and agent steps with latency, error, and token tracking
- LLM Experiments comparing prompts, models, and agent configurations
- Built-in evaluations for prompt injection plus sensitive-data scanning and redaction
- OpenTelemetry GenAI conventions support; auto-instrumentation for LangChain and OpenAI
Setup
Enable LLM Observability in your Datadog account and instrument with the Python or Node SDK, or send OpenTelemetry GenAI traces.
Prerequisites
A Datadog account; LLM or agent workloads to instrument.
Use cases
Works with
Trust score — 9/10
Scored 9/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 97; Trending score 88/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.