NVIDIA has launched Minerva, an AIOps platform that unifies telemetry from infrastructure and AI workloads to help operators find problems earlier and keep AI factories productive. The platform focuses on AI factory yield, measured through GPU hours, goodput, cluster uptime, and token production. Minerva connects signals across NVIDIA GPUs, networking, storage, schedulers, and workloads and builds a dynamic knowledge graph mapping GPUs, SuperNICs, switches, cables, and running jobs. It analyzes billions of data points per hour with supervised and unsupervised machine learning to learn normal behavior, detect anomalies, predict failures, and correlate events by topology, grouping alert storms into focused incidents with likely root causes. A native AI assistant and links to external agentic tools support natural-language investigation and recommended remediation, with optional human-in-the-loop or automated fixes via APIs.
NVIDIA has launched Minerva, an AIOps platform that unifies telemetry from infrastructure and AI workloads to help operators find problems earlier and keep AI factories productive. The platform focuses on AI factory yield, measured through GPU hours, goodput, cluster uptime, and token production. Minerva connects signals across NVIDIA GPUs, networking, storage, schedulers, and workloads and builds a dynamic knowledge graph mapping GPUs, SuperNICs, switches, cables, and running jobs. It analyzes billions of data points per hour with supervised and unsupervised machine learning to learn normal behavior, detect anomalies, predict failures, and correlate events by topology, grouping alert storms into focused incidents with likely root causes. A native AI assistant and links to external agentic tools support natural-language investigation and recommended remediation, with optional human-in-the-loop or automated fixes via APIs.