Yes. Connect the UPS USB cable to one always-on monitoring host, then publish status to the second server with Network UPS Tools or the vendor's supported network service. USB itself is not split between servers.
The monitoring host becomes a control dependency. Both servers need authenticated client configuration, shutdown policy, and enough runtime for the network path to remain available. That distinction determines the safe configuration, validation method, and rollback point.
Choose the monitor and client roles
Use the server that is most likely to remain powered and network-reachable as the UPS monitor. It owns the USB device and runs the driver, data service, and primary shutdown logic.
Configure the second server as a network client. Restrict the listener to the management network, use dedicated credentials, and firewall the service from untrusted clients.
Keep the switch between the servers on UPS power. A working monitor cannot notify a client after the network has died.
Set thresholds from runtime signals
Observe real UPS load, estimated runtime, battery age, and low-battery events with both servers active. Nameplate capacity is not a shutdown schedule.
Shut down the less critical or slower-to-stop server first, then the monitor host with enough battery reserve to finish filesystem sync and power off cleanly.
Use the table below to classify the design before relying on it.
| Observed state | Verdict | Next action |
|---|---|---|
| USB monitor plus protected switch | Network clients receive status | Recommended |
| USB cable split or switched | Unsupported ownership model | Do not use |
| Monitor reachable but guests ignore event | Incomplete shutdown chain | Add guest/VM policy |
Configure failure-safe behavior
Set clients to treat stale UPS data as a fault after a bounded interval. Decide whether they should shut down when the monitor disappears during an outage rather than wait indefinitely.
Ensure hypervisors stop guests before the host. The ZimaSpace article on UPS signals reaching hosts but not VMs explains that host notification and guest shutdown are separate problems.
Jeff Geerling's independent NUT monitoring walkthrough demonstrates one USB-connected monitor serving network clients.
Run a controlled mains-failure test
With current backups confirmed, remove wall power from the UPS—not the UPS output—and watch both servers receive the on-battery state. Restore mains before the battery is exhausted on the first test.
Then test the configured shutdown threshold, confirm the client stops first, and inspect logs on both systems after restart. Also disconnect the management network during a simulated outage to verify stale-data policy.
Proceed when both systems shut down in order and the monitor remains reachable long enough. Stop if the switch is not protected, USB ownership changes after reboot, or the client has no safe response to monitor loss.
Support & Tips
More to Read

Can You Replace a Noisy Mini PC Fan Without Changing Thermal Control?
Yes—if the replacement matches the electrical interface, airflow, and feedback signals; connector fit alone does not preserve thermal control.

Can a Home Server Resume Services in Dependency Order After UPS Recovery?
Yes—use explicit boot dependencies and readiness checks; restart policies alone do not guarantee services become usable in the right order.

Can You Use Wake-on-LAN After a Complete Power Loss?
Sometimes—WOL needs standby power and firmware/NIC state to recover after AC returns; it cannot wake a machine while power is absent.

