Release Notes for EGS Version 1.15.0
Release Date: 22nd Aug 2025
The Elastic GPU Service (EGS) platform is an innovative solution designed to optimize GPU utilization and efficiency for your AI projects. EGS leverages the power of Kubernetes to deliver optimized GPU resource management, GPU provisioning, and GPU fault identification.
We continue to add new features and enhancements to EGS.
These release notes describe the new changes and enhancements in this version.
Across our documentation, we refer to the workspace as the slice workspace. The two terms are used interchangeably. EGS Controller is also referred to as KubeSlice Controller in some diagrams and documentation.
What's New 🔈
License Management
We have introduced a new license feature that allows users to manage GPU resources more effectively. The enforcement of the license is based on the number of GPUs specified in the license file. After the licensed GPU limit is reached, no new GPU requests can be scheduled. The license count applies to the onboarded GPU resources within the Kubernetes clusters.
EGS supports both trial and enterprise licenses. To request a license, please contact the Avesha Support team with your organization's details and usage requirements. You can reach out through the Avesha EGS Registration page. For more information on how to manage license, see License Management.
GPU Provision Requests Management
The GPU Provision Request (GPR) feature has been enhanced to provide smarter automation, better flexibility, and stronger visibility into GPU allocation.
Auto GPU Selection
EGS automatically identifies and provisions GPUs that match the requested type. This reduces manual intervention, speeds up provisioning, and ensures the best-fit GPUs are allocated without conflicts.
Auto Cluster Selection (Multi-Cluster)
This enhancements extends GPR feature across multiple clusters. The system evaluates GPU availability, workload requirements, and cluster utilization to automatically place workloads in the most optimal cluster, improving efficiency and minimizing resource fragmentation.
For more information, see:
- How to create and manage GPU Requests from the Admin Portal.
- How to create and manage GPU Requests from the User Portal.
Wait Time Improvements
GPU Provisioning is now faster and more efficient. By automatically matching requests to the most suitable GPUs and clusters, GPR reduces scheduling delays and minimizes queuing. This ensures workloads start sooner, improves overall throughput, and enhances user experience in multi-tenant environments.
Enhanced Visibility and Monitoring
EGS now provides enhanced visibility and monitoring capabilities to help administrators and users better manage GPU resources.
Inventory Summary
A new consolidated view provides a cluster-wide summary of GPU resources. On the Inventory page, under Managed Nodes, you can see:
- Total number of GPUs across clusters
- GPUs currently in use and those available for allocation
- Total GPU nodes and nodes with partial allocation
This visibility helps administrators monitor resource health, identify underutilized GPUs, and plan capacity more effectively.
For more information, see:
- View GPU Inventory Summary from the Admin Portal.
- View GPU Inventory Summary from the User Portal.
Dashboard Enhancements
The dashboard has been expanded with richer GPU metrics and workload analytics. You can drill into GPU usage patterns, performance trends, and workload distribution, enabling data-driven decisions for scaling and optimization.
For more information, see:
- Monitor the Dashboard from the Admin Portal.
- Monitor the Dashboard from the User Portal.
SDK and API Enhancements
The EGS Core API and SDK has been strengthened to support deeper integrations and advanced automation use cases. API and SDK provides programmatic access to GPU provisioning, inventory, and monitoring.