NVIDIA Attribution is a workload-centric failure-attribution system used across NVIDIA AI infrastructure that correlates workload-scoped logs, metrics, telemetry, and scheduler state to return a structured diagnosis including outcome, root cause, affected nodes, supporting symptoms, and evidence trail. According to the article, a single error in distributed AI workloads can produce cascading symptoms, and the reported exception shows where the workload stopped, not where the failure began. Deterministic rules make results reproducible and customizable, and structured output enables integration with existing health and recovery systems. The open-source system has been used in production across NVIDIA AI infrastructure for years and analyzes millions of workloads each week across hundreds of thousands of GPUs. In one large RDMA-backed cluster, it correlated port errors, connection resets, and scheduler events to identify patterns around specific host and switch ports, helping teams target fabric repairs and congestion tuning. It includes two deployable components, the NVIDIA Attribution Service and the NVIDIA Attribution Blueprint, both utilizing a core attribution engine; the Service makes attribution available through a RESTAPI, while the Blueprint connects cluster schedulers and telemetry sources to the Service and writes the result. The Service analyzes one workload attempt at a time through four stages: scoping the workload, collecting observations, extracting and correlating events, and evaluating timing and interaction to separate downstream symptoms from the causing event.