We've expanded the metrics available in the Console with new per-device and per-resource breakdowns, giving you more granular visibility into both virtual machine and deployment performance.

We've completely rebuilt the Windows VM provisioning workflow for both Standalone and Cluster Hypervisors, making it faster and more flexible to deploy Windows VMs.

The new provisioning experience lets you upload a Windows ISO directly to a template and deploy VMs without any further manual intervention. Gallium automatically provisions the appropriate drivers during deployment, removing what was previously a multi-step process.

After a year of development alongside our launch customers, we're pleased to announce that both Cluster and Standalone Hypervisor deployments are now publicly available.

Second generation Hypervisor​

Both products are built on our second generation Hypervisor, which brings a broad set of new capabilities:

• PCIe Passthrough — assign physical devices directly to virtual machines
• Distributed Storage — shared storage across cluster hypervisors without external dependencies
• Live Migration — move running VMs between hypervisors with no downtime
• High Availability — automatic failover to keep workloads running if a hypervisor goes down
• Hardware Profiles — more granular control over virtual machine hardware configuration
• Real-time Metrics — live monitoring of both cluster and virtual machine resources
• Adapter Groups — active/failover network configurations for network resilience

Preview builds of the Cluster deployments are now available on request.

Cluster is built on our second generation Hypervisor, a ground-up rebuild that introduces a wide range of new capabilities. This includes PCIe passthrough, distributed storage, live migration, high availability, hardware profiles, and real-time metrics.

Multi-hypervisor architecture​

Cluster deployments support multi-node topologies with shared storage and automatic workload failover. VMs can be live migrated between nodes with no downtime, and the platform will automatically restart workloads on healthy nodes in the event of a failure.