Introducing AI Observability for OpenTelemetry

New Relic AI Observability for OpenTelemetry is presented as a platform for unifying open-source OTel and native APM telemetry to address fragmented visibility into generative AI applications, token spending, performance, and security risks. Its “Normalize on Read” approach keeps OTel data in native span tables and dynamically adapts queries at runtime, aiming to avoid data duplication, ingestion transformations, and associated storage and processing costs. The platform also supports account-wide model inventories and cross-instrumentation comparisons of cost, latency, errors, and quality, enabling teams to assess different LLM configurations. To identify “Shadow AI,” it scans incoming traces for GenAI semantic attributes and automatically tags discovered AI-enabled services, creating a current inventory without manual registration. It further routes OTel trace data into evaluation pipelines for measures such as PII masking, prompt-injection detection, and toxicity screening. New Relic states that organizations can use existing OpenTelemetry instrumentation and trace routing without custom code changes or vendor-specific SDKs.

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.