Amazon Elastic Container Service integrated NVIDIA GPU monitoring into CloudWatch Container Insights with enhanced observability. This capability exposes metrics on GPU utilization, memory, hardware health, and thermal states for GPU-accelerated containers.
The availability of detailed GPU metrics empowers users to optimize capacity, detect performance issues, and monitor hardware health proactively across ECS-managed instances running on relevant EC2 instance types, improving operational efficiency for AI/ML workloads and other GPU-intensive applications.
Source & further reading
Amazon ECS Managed Instances now supports NVIDIA GPU metrics ↗Source published 2026-04-30. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.