Allocating the appropriate amount of CPU and memory to namespaces and virtual machines is essential for achieving optimal performance and cost efficiency—especially as organizations scale across multiple clusters. Over-provisioning leads to idle resources and wasted spend, while under-provisioning risks degraded application performance. To address this, Red Hat is introducing right-sizing recommendations in technology preview as part of Red Hat Advanced Cluster Management for Kubernetes 2.15 release. This feature empowers Red Hat OpenShift Virtualization users to analyze, visualize, and optimize resource allocation across their environments using built-in observability and analytics.
Why right-sizing matters
In multicluster, hybrid-cloud environments, managing resource efficiency across thousands of workloads—spanning containers and virtual machines is one of the hardest operational challenges. While several tools in the market provide sizing or cost insights at an application or node level, Red Hat Advanced Cluster Management right-sizing offers something fundamentally different: end-to-end observability and recommendations across Cluster → Namespace → Virtual Machine layers, all natively integrated into OpenShift and Prometheus-based telemetry.
Red Hat Advanced Cluster Management right-sizing helps to:
- Optimize resource utilization: Reduce idle resource allocations and maximize cluster efficiency.
- Enhance cost efficiency: Avoid over-provisioning and lower operational expenses.
- Improve performance and predictability: Prevent under-provisioning that can lead to latency and instability.
Right-sizing in action
The technology preview includes highlights for OpenShift Virtualization workloads. The right-sizing capability leverages Prometheus recording rules and Grafana dashboards integrated into the Red Hat Advanced Cluster Management console to provide actionable recommendations.
Two primary dashboards are available:
Cluster & namespace right-sizing dashboard
Released in July 2024 (technology preview with version 2.14), this view helps platform engineers evaluate aggregate CPU and memory utilization across clusters and namespaces.
Virtual machine level right-sizing dashboard
Introduced in March 2025 (developer preview), this dashboard extends visibility to OpenShift Virtualization workloads, helping administrators identify over- and under-provisioned VMs at a glance.
Navigating the Grafana dashboards
Each right-sizing dashboard includes multiple perspectives, including:
- Total overestimation and underestimation: Aggregated CPU/memory deltas across all VMs in a cluster.
- Per-VM utilization tables: View detailed CPU and memory over/underestimation per virtual machine.
- Time-series trends: Visualize how resource usage evolves over time for a selected VM.
By clicking on any VM name, users can explore detailed historical utilization charts, enabling data-driven tuning decisions.
The following video is demonstration of this feature.
Prerequisites
Before using the technology preview, ensure the following prerequisites:
- Red Hat Advanced Cluster Management for Kubernetes (latest release)
- OpenShift Virtualization enabled
- Observability components deployed and configured (Prometheus, Thanos, Grafana)
For full installation steps and known limitations, refer to the official product documentation.
What’s next
The Red Hat Observability and Advanced Cluster Management engineering teams are actively working toward the GA release of right-sizing recommendation feature.
We welcome your feedback. Try the technology preview and share your thoughts through the Red Hat OpenShift feedback form. Your input will shape the path to full productization.
