Avesha Resources / Blogs
Veena Jayaram
Staff Technical Writer/Program Manager

We're excited to announce the release of Smart Scaler version 2.17.0, our latest major update that is live as of Dec 8, 2025 and a later update that is live as of Jan 20, 2026. The major and update version releases are aimed at enhancing predictive scaling capabilities in Kubernetes environments. This version introduces new features and improvements designed to optimize resource utilization, reduce costs, and provide better visibility into your cluster operations.
We’re also glad to announce the release of Smart Scaler Agent version 2.9.40.
Smart Scaler is an enterprise-grade solution designed to address the complexities of Kubernetes scaling and resource management. It provides intelligent autoscaling capabilities that go beyond traditional Horizontal Pod Autoscaler (HPA) by considering multiple metrics simultaneously and making scaling decisions based on advanced algorithms.
Smart Scaler comes with Smart Sizing, which involves adjusting resource requests and limits to ensure your applications have the optimal amount of resources to operate efficiently. This helps prevent overprovisioning, which can unnecessarily increase costs, and underprovisioning, which may lead to performance issues.
Smart Scaler now includes Ask Smart Scaler, a stabilized feature that allows you to enter natural language queries about your system and receive data insights or visualizations in return. This unlocks powerful ways to explore utilization trends without manual drilling.
For more information, see Ask Smart Scaler.
We’ve added a brand-new Metrics Version tab in the dashboard. This gives you the ability to analyze CPU and memory request statistics across clusters or drill down into a specific cluster’s performance metrics.
For more information, see Explore Metrics Insights.
Smart Sizer has been redesigned with improved visibility:
For more information, see Explore Metrics Insights.
Now you can revisit and reuse previous Ask Smart Scaler queries through a chat history feature, helping you quickly rerun common or complex queries without retyping them.
For more information, see Use the Chat History.
We’ve refined the dashboard layout and interactions to be more responsive and intuitive. The result is a smoother, more coherent dashboard experience that lets you explore scaling insights with greater ease.
For more information, see Dashboard.
To make sizing insights more actionable, the Degree of Provisioning parameter on the Smart Sizing page has been renamed to Efficiency, aligning better with the intent of the metric.
For more information, see Monitor Smart Sizing Recommendations for Application Resources.
The standout feature in this release is High Availability (HA) support in the Smart Scaler Agent. With HA enabled, the agent now coordinates priority-based custom metrics across multiple agent instances. This ensures that the Horizontal Pod Autoscaler (HPA) always receives the highest-quality metrics, improving stability, accuracy, and resilience of scaling decisions.
For more information, see Configure High Availability for the Inference Agent.
Existing Smart Scaler users can upgrade to version 2.17.0 and agent version 2.9.40 or later through our standard upgrade process. New users can install our Smart Scaler agent using Helm.
For detailed installation instructions and documentation, visit our Smart Scaler documentation.
We encourage all users to upgrade to the latest version to take advantage of these new features and improvements.
Copied