New Relic announced that New Relic AI monitoring (AIM) is now integrated with Amazon Bedrock, a fully managed service by Amazon Web Services that makes foundation models (FMs) from leading AI companies accessible via an API to build and scale generative AI applications.

AWS customers can now use New Relic to gain greater visibility and insights across the AI stack, making it easier to troubleshoot and optimize their applications for performance, quality, and cost.

While AI is revolutionizing modern applications, it introduces new challenges and complexity to organization’s tech stacks. AI tech stacks include new components like large language models (LLMs) and vector data stores and generate additional telemetry to track such as quality and cost. AIM solves these new challenges by bringing APM to the AI stack. Similar to how engineers monitor their application stack with New Relic APM, AIM provides engineers with full visibility into all components of the AI stack. AIM provides a single easy view to troubleshoot, compare, and optimize different LLM prompts and responses for performance, cost and tokens, and quality issues including hallucinations, bias, toxicity, and fairness across all models supported by Amazon Bedrock.

AIM integrates with Amazon Bedrock to provide in-depth end-to-end observability. With AIM’s built-in integrations such as Langchain, Amazon Bedrock customers can get metrics and tracing throughout the life-cycle of LLM prompt and response, ranging from raw prompts to repaired and business-compliant responses.

Key features and use cases include:

- Auto instrumentation: New Relic agents come equipped with all AIM capabilities, including full AI stack visibility, response tracing, model comparison, and more for quick and easy setup.

- Full AI stack visibility: Holistic view across the application, infrastructure, and the AI layer, including AI metrics like response quality and tokens alongside APM golden signals.

- Deep trace insights for every LLM response: Trace the lifecycle of complex LLM responses built with tools like LangChain to fix performance issues and quality problems such as bias, toxicity, and hallucination.

- Compare performance and costs: Track usage, performance, quality, and cost across all models in a single view; optimize use with insights on frequently asked prompts, chain of thought, and prompt templates and caches.

- Enable responsible use of AI: Ensure safe and responsible AI use by verifying that responses are appropriately tagged to indicate AI-generated and are free from bias, toxicity, and hallucinations using response trace insights.

- Instantly monitor your AI ecosystem: The most comprehensive solution for monitoring the entire stack of any AI ecosystem with 50+ integrations and quickstarts including:
Orchestration framework: Langchain
Vector databases: Pinecone, Weaviate, Milvus, FAISS, Zilliz
LLM: Amazon Bedrock (models from AI21 Labs, Amazon, Anthropic, and Cohere)
AI infrastructure: Amazon SageMaker

“AI workloads are now part of modern organizations’ application architectures, and observability is essential for any company building AI applications,” said New Relic Chief Product Officer Manav Khurana. “Today’s news builds upon our deep work with AWS to bring the power of observability to engineers and developers who are modernizing their tech stacks. And by putting our AI monitoring solution front and center with AWS customers, we are multiplying our ability to reach every engineer using leading LLMs like Anthropic.”