Harvester v1.8.0 UI now proactively alerts users when a Virtual Machine requires a restart to fully apply configuration changes, such as interface binding updates or specific CPU/memory modifications. The dashboard continuously polls the VM's status conditions and displays a "Pending Restart" indicator when necessary. This prevents unapplied changes from silently blocking future live updates or resource hotplug operations, ensuring users can promptly resolve the state using the standard UI restart action.

Harvester v1.8.0 addresses an issue where live migrating VMs between hosts with different CPU generations (for example, AMD Ryzen 5th Gen to 3rd Gen) would fail due to strict CPU feature-matching checks by KubeVirt/libvirt. Users can now explicitly define a specific CPU model (for example, IvyBridge, host-model, or default) for a VM during creation via the Harvester UI or Terraform provider. This ensures the VM is only scheduled and migrated across nodes that support the selected baseline CPU model, enabling reliable backward compatibility and successful live migrations in mixed-hardware clusters.

In Harvester v1.8.0, a dynamic visual topology diagram is added to the VPC detail page, providing a clear representation of logical switches and network relationships. The interactive map displays the complete hierarchy from VPCs and Subnets to Overlay Networks and individual VMs while accurately reflecting VM power states by graying out stopped or paused instances. Users can easily navigate complex network architectures using pan and zoom controls, legend toggles for filtering node types, click-to-highlight node relationships, and clickable VPC peering links that seamlessly redirect to associated topology pages.

Harvester v1.8.0 decouples the upgrade logic from the core harvester binary into a stand-alone upgrade-manager. This architectural change enables faster delivery and testing of upgrade-related features and bug fixes without waiting for full release cycles. The upgrade lifecycle is now efficiently managed by `components—upgrade-shim` (early stage preparation), `upgrade-repo` (metadata and image hub), and `upgrade-manager` (custom lifecycle controllers) improving the overall maintainability and upgrade reliability.