Content Deep Dive

Datadog LLM Observability: Features, Pricing, and Alternatives

Blog post from New Relic

Post Details
Company
Date Published
Author
John Blust
Word Count
2,084
Company Posts That Month
12
Language
English
Hacker News Points
-
Post removed?
No
Summary

Datadog LLM Observability, now branded Agent Observability, monitors and troubleshoots LLM applications and AI agents through end-to-end tracing, quality and security evaluations, prompt experiments, and token, latency, and cost tracking. Its traces cover model calls and agent workflows, while evaluations can flag issues such as hallucinations, prompt injection, and sensitive-data exposure. As of November 2026, the standalone product offers a free plan with 40,000 billable LLM spans per month and a Pro plan starting at $160 per month with annual billing; each model call counts as a span, while tool, retrieval, and workflow spans do not. Both plans include core observability and experimentation features, with longer retention available at added cost. Teams should weigh span-based pricing, supported instrumentation, and whether they want AI monitoring separate from or integrated with existing APM. Alternatives include New Relic AI Monitoring, which connects AI telemetry to existing APM, Langfuse, which offers a self-hostable core, LangSmith, which focuses on agent tracing and evaluation, and Arize Phoenix, which supports OpenTelemetry-based and local workflows. The article emphasizes that diagnosing AI application issues often requires examining model behavior alongside the services, databases, and infrastructure involved in each request.

Trends Found in this Post
Trend Post Mentions Total Month Mentions Posts Companies MoM
LLM 54 3,674 823 199 -44%
Observability 49 2,390 523 156 -23%
AI Agents 12 6,233 1,230 245 -11%
OpenTelemetry 6 822 142 56 -4%
MCP 2 9,757 959 203 -18%
Vector Search 2 1,251 223 92 -37%
Harness engineering 1 122 76 53 -20%
Use This Data

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.