<div>
<p>The article introduces a new feature in Amazon CloudWatch Container Insights with Enhanced Observability for EKS that provides auto-discovery and monitoring of critical health and performance metrics from NVIDIA GPUs. This allows users to easily understand the health and performance of their GPU-accelerated workloads, such as AI/ML training jobs, and troubleshoot issues more quickly.</p>
<p>Specifically, the article covers:</p>
<ul>
<li>Overview of the new GPU metric monitoring capabilities in CloudWatch Container Insights</li>
<li>Benefits of the enhanced observability, including automatic dashboards, faster problem isolation, and resource optimization</li>
<li>How to get started by enabling the feature in the EKS console or CloudWatch Agent configuration</li>
<li>Availability of the NVIDIA GPU metrics in all AWS regions, including pricing information</li>
<li>Link to the Container Insights user guide for further details</li>
</ul>
</div>


CloudWatch Container Insights now delivers observability for NVIDIA GPUs on EKS

Related articles