Changes in Release 1.130 (2026/09/27)

Partner Portal

Fixed: An expansion invoice could print a line item that did not add up to its own total.
Fixed: Invoices, credit notes and quotes applied the partner's current discounts to a total that had already been frozen, so a document could contradict the offer it was built from.

New: A High Resource Usage card on the Dashboard lists assets and clients whose CPU or memory has stayed above 95% for five minutes, with a matching notification that clears once the device has been back under the threshold for three minutes.
New: Client Resources events record every time a device crosses the CPU or memory threshold, and can be forwarded like any other event class.
New: Virtual assets and subnets can be cloned, carrying over their services, tags and policy assignments.
Changed: Policies can be filtered by Has Notification, like assets and clients.
Changed: Dashboard cards pack without leaving half-empty slots between them.

Changed: Clients behind the same firewall share one NAT classification, so a large site behind a single public IP is classified much faster.
Fixed: An asset with several hundred peers could fail to connect to a relay at all, retrying every few seconds and staying unreachable.
Fixed: At a large site behind one public IP address, clients could fail classification and reconnect in a loop that never recovered on its own.
Fixed: On Windows and macOS, the agent's section of the hosts file could be duplicated over and over, and the stale entries left behind could keep a name resolving to an old address.
Fixed: On Windows, an automatic update that ran into a system shutdown could leave the agent installed but with none of its services registered.