NVIDIA is helping enterprises build AI factories that are cost effective, scalable and high-performing — equipping them to meet the next industrial revolution. The design for an Enterprise AI Factory integrates seamlessly with enterprise systems, data sources, and security infrastructure through NVIDIA’s partner solutions.
As described in the NVIDIA Enterprise AI Factory - Design Guide, it is with Observability that teams can understand the operation of long-running agents, extensive telemetry is necessary to observe the internal reasoning of the system. AgentOps addresses this requirement through a three-pronged monitoring strategy: traces detailing the workflow, logging user activity and responses of the models, and key performance metrics for the outcomes. This comprehensive set of signals provides a complete view, illustrating the entire execution path and quantifying the business value derived from these long-duration tasks.
The NVIDIA AI Enterprise Infrastructure software encompasses all necessary components for managing and optimizing infrastructure along with AI workloads. NVIDIA provides Release Branches to meet organizational needs. The NVIDIA Kubernetes Operators facilitate a standardized management of NVIDIA GPUs, AI models, and network resources within Kubernetes environments as well as standalone Docker on Red Hat Enterprise Linux, SUSE Linux Enterprise Server, and Ubuntu.
As a NVIDIA validated partner, Dynatrace serves as a critial part of AI factorys by bringing end-to-end and full stack observability for LLMs and Agentic Generative AI applications running on NVIDIA GPUs and NVIDIA AI Enterprise that brings together microservices, frameworks, and libraries for AI development with advanced GPU orchestration and infrastructure management in a fully supported, production-ready, commercial software suite.
This repo provides various guides to show how easy it is to setup and monitor the key tiers and components of the NVIDIA enterprise stack, LLMs, and Generative AI applications with Dynatrace. Guides are structured according the visual shown below.
- End-to-End Application Observability
- NVIDIA NIM, NeMo and other microsevices
- AI application SDKs, frameworks and libraries
- Kubernetes
- GPU telemetry
Note These guides are not officially supported by Dynatrace, but the integrations themselves are.
