Network UPS Tools with Home Assistant and NAS: Safe Automatic Shutdown

Configure Network UPS Tools (NUT) for Home Assistant, Synology, UGREEN, QNAP and Proxmox, with shutdown clients, safe testing and recovery.

Quick Summary (TL;DR):
Network UPS Tools (NUT) lets one supported UPS protect multiple machines, not just the NASPriceAliExpress price or mini PCPriceAliExpress price attached to its USB lead. A NUT driver reads the UPS; the upsd server shares its state on the local network (normally TCP port 3493); upsmon clients request orderly shutdown when power becomes critical. For a mixed home lab, choose one reliable, UPS-powered NUT server, keep the Ethernet switch powered by the same UPS, configure each Linux or Proxmox host as a genuine shutdown client, and use Home Assistant primarily for monitoring and alerts. Synology DSM, QNAP QTS and UGREEN UGOS Pro also offer built-in UPS sharing, but their network client support and firmware options differ. Do not assume a Home Assistant dashboard integration by itself shuts down the host. Choose shutdown thresholds from measured runtime, test the entire power-failure and recovery path, and do not expose NUT port 3493 to the internet. This guide uses documented interfaces and configuration examples; no UPS or NASAliExpress price hardware was tested by esp32.co.uk/.

Why Network UPS Tools Matters for a Home Lab

A UPS supplies temporary electrical power. That is only half of a storage-protection strategy. If your NAS, Proxmox server and Home Assistant machine continue working until the battery empties, they can still lose power during disk writes. Automatic shutdown converts remaining battery time into a controlled stop of applications, virtual machines and storage pools.

A typical USB UPS has just one management connection. NUT makes that single connection useful to several systems by separating UPS telemetry from local shutdown decisions. Every participating computer needs an instruction path to the UPS status, and every important network component on that path must remain powered during an outage.

You do not need to write elaborate Home Assistant automations to achieve this. On Linux and Proxmox, upsmon is designed to handle the critical shutdown path; the Home Assistant NUT integration is valuable for status, notifications and dashboards. On a vendor NAS, start with its built-in, supported UPS features.

How NUT Works: Driver, Server and Shutdown Clients

ComponentRoleWhere it normally runs
NUT device driverReads UPS status such as on-battery (OB), low-battery (LB), charge and estimated runtimeMachine directly attached to UPS USB; or a supported SNMP/network driver host
upsdPublishes UPS information to authenticated or permitted NUT clientsNUT server, on backed-up power
upsmon primaryMonitors power and coordinates forced shutdown; may command a final UPS output shutdown if configuredSystem responsible for managing the UPS
upsmon secondaryFollows NUT power status and safely shuts down its own machineProxmox nodes, Linux home servers, compatible clients
Home Assistant NUT integrationReads published values for sensors, alerts and supported actionsHome Assistant instance with network access to NUT server

Modern NUT documentation uses primary and secondary roles; older tutorials may call these master and slave. These terms describe permission and shutdown coordination, not which machine is physically larger. The server daemon upsd and monitoring daemon upsmon do different jobs even when they run on the same computer.

Mains electricity
       |
   [UPS output]---- NAS power input
       |---------- Proxmox / mini PC power input
       |---------- Router + Ethernet switch power
       |
 UPS USB data cable
       |
 [NUT server: driver + upsd + upsmon primary]
       |
       +---- LAN ---- [Proxmox: upsmon secondary]
       +---- LAN ---- [Linux home server: upsmon secondary]
       +---- LAN ---- [Home Assistant: NUT integration for alerts]

The network cables can be on ordinary Ethernet; they do not need any special UPS data wire. The UPS must power the Ethernet switch while clients still need it. If the switch loses mains first, a remote client may never receive the low-battery warning.

Choose the Right NUT Architecture Before Installing Anything

SituationPractical designWatch for
One NAS onlyUSB UPS directly to that NAS; use built-in power settingsNo need for NUT network clients unless other devices require status
Synology plus other Synology NAS devicesDSM UPS server with permitted Synology clientsSupported model/DSM version and protected switch
QNAP NAS pairUSB-connected QNAP as network UPS master; second QNAP as slaveSame subnet and supported QTS settings
UGREEN NASync pairUGOS Pro NUT server with compatible client NASFirmware and interface names vary
NAS + Linux/Proxmox machinesDedicated supported NUT server + Linux upsmon clientsVendor-specific interoperability is not guaranteed
Home Assistant monitoring onlyExisting NUT server + Home Assistant NUT integrationIntegration does not install a UPS driver or guarantee host shutdown

Our preferred mixed-vendor design: if you already own a reliable always-on Debian or Proxmox host, configure a supported USB UPS under its control and let the other Linux machines act as clients. If the NAS should be the NUT server, confirm its firmware actually provides a compatible server endpoint and client credentials. Avoid combining undocumented NAS SSH tweaks with automatic shutdown of irreplaceable storage.

A small dedicated Raspberry Pi can act as a NUT server, but it is another device you must power, maintain and back up. A single-board computer relying on an SD cardPrice also adds failure modes. It is not automatically better than the server you already have.

Check UPS Compatibility First

NUT can read many USB HID UPS devices, including supported models from APC, CyberPower and Eaton. However, two UPS models from the same manufacturer can use different monitoring protocols, and the fields they expose differ. Before purchasing, consult the official NUT hardware compatibility list and the documentation for your precise regional model.

  • Confirm the exact product identifier and connection type: USB HID, USB vendor protocol, serial, or network management card.
  • Confirm the specific driver. usbhid-ups is common for compliant HID devices but is not universal.
  • Check that ups.status reports both mains restoration (OL) and battery operation (OB) correctly.
  • Do not insist on percentage and runtime fields if the hardware only reports basic state. Reliable OB/LB handling matters more than pretty dashboards.
  • Verify battery-backed sockets, output watts, runtime and replacement battery; USB compatibility does not make an undersized UPS adequate.

For hardware choices, see Best UPS for Home Assistant, Router and NAS and the detailed UPS VA vs Watts sizing guide.

Build a Safe Shutdown Timeline, Not a Guess Based on VA

The right delay depends on the actual watt load and the time required for every machine to stop safely. You need enough capacity to survive the triggering delay plus the slowest VM shutdown plus the final server shutdown and some contingency. A 1000VA label does not tell you those numbers.

EventIllustrative actionWhy it matters
0 seconds: mains failsUPS switches to battery; NUT status changes to OBStart monitoring, not necessarily immediate shutdown
30–60 secondsOptional notification or early VM workload reductionAvoid unnecessary shutdown from brief power dips
When chosen safe threshold is reachedInitiate orderly shutdown for Proxmox, NAS and any servicesDo not wait for battery to reach 0%
Before expected battery exhaustionGuests finish; hosts unmount storage and haltLeave an allowance for cold/aged batteries
Power returnsInspect UPS output, server auto-start and application healthA clean shutdown is only half of successful recovery

These times are planning examples only, not recommended universal timer values. Some NAS devices can take several minutes to stop disks and containers. A busy Proxmox host with unresponsive VMs may take much longer. Time the actual behaviour using a controlled maintenance shutdown before setting battery thresholds.

The official NUT low-battery (LB) condition can trigger forced shutdown, but the UPS decides when it enters that state. A system that requires longer shutdown lead time may need a carefully designed earlier trigger using upssched or supported NAS settings. Set the threshold conservatively, then verify it physically.

Install NUT on a Debian or Ubuntu Server

The following is an illustrative NUT server configuration for a single supported USB HID UPS on an always-on Debian-family host. Device names, service names, package file locations and permission models vary between distributions and versions. Record your existing configuration and check the official manual pages before applying anything to production.

sudo apt update
sudo apt install nut

# Identify attached USB devices
lsusb

# After configuration and service startup:
upsc labups@localhost ups.status

The sample uses labups as a fictional UPS name and 192.168.10.5 as a fictional NUT server IP. Substitute your own static or reserved address; do not copy these values without changing them.

Set the NUT Role in nut.conf

On Debian-family packaging, the mode generally lives in /etc/nut/nut.conf. Select a combined network server role for the host physically reading the UPS. For a remote client, use MODE=netclient instead. Verify the packaging in use.

# /etc/nut/nut.conf on the USB-connected host
MODE=netserver

Configure the UPS Driver in ups.conf

A single supported USB HID device often needs only the following configuration. NUT drivers for other UPS protocols may need a different driver or device port. NUT requires a port entry even though USB HID implementations commonly ignore its value.

# /etc/nut/ups.conf
[labups]
    driver = usbhid-ups
    port = auto
    desc = "Home-lab backup UPS"

If detection fails, identify the correct driver using the compatibility list instead of guessing USB vendor IDs. Make sure only one service owns the UPS USB interface; two competing drivers can cause intermittent failures.

Make upsd Available to Your LAN — but Not the Internet

By default, a NUT server may listen only on loopback. On the local server, an explicit LAN LISTEN address allows clients to connect. Keep the port behind your trusted firewall; never port-forward it on the router. NUT communication and credentials must be treated as sensitive administrative traffic.

# /etc/nut/upsd.conf
LISTEN 127.0.0.1 3493
LISTEN 192.168.10.5 3493

Allow TCP 3493 only from approved local machines and management VLANs. If your server address changes, clients can lose the UPS connection. Assign a DHCP reservation or static address and document it. Some systems require a full restart of upsd after changing its listen address.

Create Separate Monitoring Accounts

Use distinct credentials for the local primary monitor and the remote shutdown clients. Below, REPLACE_WITH_LONG_RANDOM_SECRET is a placeholder—not a usable password. Store the generated values securely, limit file permissions and never paste them into public logs or screenshots.

# /etc/nut/upsd.users
[primarymon]
    password = REPLACE_WITH_LONG_RANDOM_SECRET_A
    upsmon primary

[secondarymon]
    password = REPLACE_WITH_LONG_RANDOM_SECRET_B
    upsmon secondary

The upsmon primary and upsmon secondary directives grant purpose-specific NUT privileges. Do not grant all instant commands to a Home Assistant dashboard account merely to display UPS charge. If monitoring without authentication is supported by your chosen NUT setup, restrict read access at the network layer.

Configure the Primary Monitor

The USB-connected host needs its own power-failure response. In /etc/nut/upsmon.conf, set a primary monitor entry with the actual credential you created. The rest of this configuration file contains many important defaults; edit deliberately rather than blindly replacing the entire file.

# Relevant line in /etc/nut/upsmon.conf on server
MONITOR labups@localhost 1 primarymon REPLACE_WITH_LONG_RANDOM_SECRET_A primary

# Example concept; confirm the actual shutdown command
SHUTDOWNCMD "/sbin/shutdown -h now"

The 1 indicates one power-supply equivalent required from this UPS. NUT supports more complex topologies, but for a machine with one PSU supplied by one UPS, that is the usual design. Do not copy advanced MINSUPPLIES, HOSTSYNC or FINALDELAY values from unrelated blog posts without understanding the full sequence.

Start the Services and Verify Telemetry

Systemd unit names depend on your Linux distribution and package version. Debian/Ubuntu commonly use units such as nut-server and nut-monitor plus a driver service/enumerator. Check the installed units with systemctl list-unit-files "nut*"; do not assume the server is running merely because apt install completed.

systemctl list-unit-files "nut*"
systemctl status nut-server
systemctl status nut-monitor

# Observe current values; this is read-only
upsc labups@localhost ups.status
upsc labups@localhost battery.charge
upsc labups@localhost battery.runtime

The last two fields are optional and may report unknown variable if the UPS does not supply them. A healthy mains-powered device commonly reports OL. When appropriately tested with input mains removed and the load still on UPS output, it should report OB. If it remains permanently OL, fix the driver or configuration before trusting automatic shutdown.

Add a Proxmox VE or Other Linux Client

On a Proxmox VE host, install the client-side NUT components through the supported Debian packages. The host should watch the server without competing for the USB cable. Keep the VM shutdown responsibility at the hypervisor level rather than trying to make every container issue its own poweroff.

sudo apt install nut-client

# /etc/nut/nut.conf
MODE=netclient

# Relevant line in /etc/nut/upsmon.conf
MONITOR labups@192.168.10.5 1 secondarymon REPLACE_WITH_LONG_RANDOM_SECRET_B secondary

# Inspect service names on your system
systemctl list-unit-files "nut*"

# Verify client-to-server read access
upsc labups@192.168.10.5 ups.status

For a remote secondary, set a suitable local SHUTDOWNCMD and verify that nut-monitor runs at boot. Network reachability alone is not enough: upsc can read status while upsmon is misconfigured or stopped. Confirm both the monitoring daemon and its shutdown actions in the logs.

Proxmox host shutdown normally initiates orderly stop of hosted guests, but final success depends on guest agents, ACPI handling, stop/shutdown ordering and the guest timeout. Ensure important NAS-backed VM disks remain accessible until the VMs have halted. If the NAS shuts down first, even a perfectly configured NUT client on Proxmox may not prevent data errors.

Special Case: Proxmox, NAS and Home Assistant Share One UPS

A robust operating order often looks like this: the hypervisor receives the critical power warning and shuts down its Home Assistant VM and other services; a physical NAS that provides storage remains up until its active clients release network files; the NUT primary is the final machine to halt and, where supported, the UPS subsequently powers down its load.

However, a built-in vendor NAS timer may not wait for Proxmox to complete guest shutdown, and the primary NUT server usually waits only a limited time for secondary clients before continuing. There is no universal shutdown order enforced across heterogeneous NAS and hypervisor platforms. For complicated dependencies, design an explicit and tested sequence rather than assuming every device reacts to LB at the same time.

  • Identify storage dependencies: VMs on NFS/iSCSI NAS, container volumes, databases and Home Assistant recorders.
  • Measure guest shutdown time during a normal maintenance window and adjust the Proxmox startup/shutdown order and timeout accordingly.
  • Keep NAS/network equipment running until required filesystems are no longer mounted by guests.
  • Reserve sufficient UPS runtime after the longest shutdown path; an early planned halt can be safer than waiting for low battery.
  • Document a manual recovery order for the first restart following a long outage.

Connect Home Assistant to NUT for Monitoring and Alerts

Home Assistant provides a built-in Network UPS Tools integration at Settings → Devices & services → Add Integration → Network UPS Tools (NUT). It connects to an existing NUT server; it does not speak directly to a UPS over USB. Enter the stable server IP/name and port 3493, and credentials if the server requires them.

Depending on what the UPS exposes, you can see status, battery charge, estimated runtime, power load and other diagnostics. Official Home Assistant documentation notes that diagnostic entities are often disabled by default to reduce storage overhead. If a desired field is absent, first check whether the NUT driver actually publishes that variable with upsc.

Use a simple alert automation for the OB condition and a second warning for a low estimated charge or runtime. Prefer the status data provided by your integration or inspect the actual friendly status options; never blindly copy an entity ID from a screenshot. In the UI, select the actual UPS entity installed on your system.

The Home Assistant integration polls the NUT server (the official documentation currently describes a default 60-second interval). This is reasonable for a dashboard, but not a substitute for a dedicated shutdown daemon with a carefully tested critical path. The published docs also distinguish sensors from commands: users need explicit NUT privileges before supported actions and switches appear.

Should Home Assistant Shut Itself Down?

For Home Assistant OS running on its own mini PC or Raspberry Pi, a supervised host shutdown is different from restarting Home Assistant Core. The documented Home Assistant CLI includes ha host shutdown, but the available integration alone does not schedule or guarantee that command is executed when batteries run low. A standalone HAOS shutdown solution must be supported by the installed host/add-on arrangement, or by a separate reliable external controller.

If Home Assistant is a Proxmox VM, the preferred design is for the Proxmox host to shut down the guest using its own managed power sequence. Home Assistant notifications then become supplemental. If Home Assistant is the only system capable of telling itself to shut down, check that the NUT server, HA automations and required integration are still functioning under every expected failure mode.

Avoid granting a broad remote UPS control account to Home Assistant solely for alerts. Power-off commands can disrupt an entire NAS array or home network, so monitor-only access is preferable until a specific control function is needed and reviewed.

Synology DSM: USB UPS and Network UPS Server

Synology DSM has built-in UPS configuration under Control Panel → Hardware & Power → UPS on supported systems. With a compatible USB UPS, enable UPS support and select the response to battery operation. The relevant protection state is commonly called Safe Mode; DSM stops services and safeguards mounted storage according to its documented behaviour.

Where multiple Synology NAS systems share one UPS, the directly connected NAS can enable network UPS server, then list the permitted Synology NAS devices by IP. Each client enables UPS support, selects the Synology UPS server type and enters the server address. Synology explicitly advises that the network switch must also be UPS-powered.

Caution: that supported DSM workflow is primarily documented for compatible Synology clients. Do not assume every generic Linux NUT client can use the Synology service, account permissions or shutdown semantics without testing the precise DSM version. Some community configurations work, but unsupported changes could be lost on firmware updates. Refer to the Synology DSM UPS help for your installed version.

QNAP QTS: Network UPS Master, Slave and SNMP

QNAP QTS 5.2 documentation locates UPS controls under Control Panel → System → External Device → UPS. A supported USB UPS can be configured to power off the NAS after a delay or enter auto-protection mode, which stops services and unmounts volumes. The USB-connected NAS can optionally enable network UPS master and specify other NAS IP addresses; clients select Network UPS slave.

An appropriately equipped UPS network card may instead communicate via SNMP, but SNMP is not the same as NUT upsd. QNAP network UPS master mode is documented for supported QNAP clients and QNAP recommends testing each slave. Read the QTS/QuTS hero instructions for your firmware before mixing QNAP with another vendor.

When choosing between power-off and auto-protection, decide whether you prioritise a fully powered-down system or quick resumption after brief outages. Test both the outage and the return of mains: automatic restart behaviour differs from entering a protected low-power state. Official details: QNAP QTS UPS configuration.

UGREEN NASync / UGOS Pro: NUT Server and Client

UGREEN now documents a NUT server feature in UGOS Pro for multi-NAS UPS monitoring, including a directly connected primary NAS and a network client NAS. Its support article states that interface screenshots depend on firmware and Control Panel version; the published example identifies version 1.6.1.2846. Verify your own firmware before following menu steps.

The supported pattern is a compatible UPS connected by USB to the primary UGREEN NAS, NUT server sharing enabled, then another NAS pointed at the server IP and port 3493 through its network UPS client settings. Set a fixed/reserved IP, protect the switch and verify that both machines shut down or enter the intended protection mode.

The UGREEN documentation focuses on NAS-to-NAS operation. General third-party NUT clients may require compatible authentication and permissions. Rather than promising cross-vendor operation for every current UGREEN firmware, validate the exact combination before relying on it. Sources: UGREEN UPS/NUT server support and UGREEN NAS NUT setup guide.

Shutdown vs Safe Mode vs Loss of UPS Output

Event or settingMeaningPotential trap
Shutdown hostOperating system stops services and powers itself offIt may not restart if the UPS output never switches off
NAS Safe Mode / auto-protectionNAS protects data and may stop/unmount storageNot always the same as full electrical power-off
NUT FSD flagForced shutdown coordination between NUT monitorsNot itself proof the UPS output was cut
UPS output shutdownUPS switches off battery-backed outputs after coordinated stopIf triggered too early, storage suffers abrupt loss of power
Return of mainsUPS resumes supply if configuredHost BIOS/UEFI and UPS behaviour both control restart

NUT uses a POWERDOWNFLAG and platform shutdown scripting for some primary setups to tell the UPS to switch off its output after filesystems have been made safe. Distribution-specific shutdown integration matters. Never copy a command such as upsdrvctl shutdown into an early outage notification hook without understanding what it will turn off.

A host with BIOS/UEFI “restore after AC power loss” enabled will usually restart only after it actually experiences a power-loss/power-restore event. If the host shuts down gracefully while the UPS continues to supply AC, it can remain powered off indefinitely. Test this recovery path deliberately, and document what will require a manual button press.

A Safe Test Plan: Prove Detection, Shutdown and Recovery

The point of a UPS test is not simply to make the dashboard say “On Battery”. You must verify that the storage stays consistent after every relevant device stops and that the system eventually restarts as expected. Schedule the test during a maintenance window with recent verified backups.

  1. Before testing: record running wattage, current UPS charge, approximate remaining runtime and which sockets supply each device.
  2. Check visibility: on the server and every client, run read-only upsc labups@SERVER_IP ups.status and check the respective NUT/OS logs.
  3. Rehearse ordinary shutdown: halt Proxmox guests and NAS services through their supported maintenance controls; measure duration.
  4. Simulate battery: where safe for the UPS manufacturer model, unplug only the UPS mains input while leaving loads connected to its battery-backed outlets. Watch for OB. Do not pull the NAS or server power lead.
  5. Observe recovery from a short interruption: reconnect mains before a shutdown threshold and confirm a return to OL without unnecessary disruption.
  6. Conduct a supervised long-outage drill: with spare runtime and physical access, let the configured sequence reach its threshold and verify each guest and host stops correctly.
  7. Validate final restart: inspect UPS output restart, BIOS settings, router, switch, NAS storage health, Proxmox guests and Home Assistant.
  8. Review logs: look for missing alarms, NUT timeouts, an early NAS shutdown, stalled guests and services that did not recover.

NUT documentation includes the upsmon -c fsd command for simulating a forced shutdown. Do not use it casually: it is intentionally capable of initiating a real coordinated shutdown. If you need a no-disruption dry run, design a controlled lab with a harmless SHUTDOWNCMD or other supported simulation under an administrator who understands the consequences. Never treat an untested FSD in production as a harmless status check.

Troubleshooting: What Common NUT Errors Actually Mean

SymptomLikely causeTargeted check
Connection refusedServer not running or only bound to loopbackInspect nut-server, upsd.conf LISTEN and TCP 3493 firewall
Unknown UPSWrong device name or driver not registeredCheck ups.conf stanza and upsc -l
Data staleUSB polling/driver or device disconnectionInspect driver logs, USB permissions and cable stability
Access deniedClient credentials or server permissions wrongVerify upsd.users and MONITOR account role
Home Assistant sensors show unavailableNUT host/IP changed, network lost or data field absentPing NUT server, query upsc, inspect entity diagnostics
NAS never shuts downDelay too long or vendor UPS mode not selectedCheck NAS settings and verify supported shutdown behaviour
Proxmox host halts but guests misbehaveGuest shutdown/agent timeout or NAS storage disappeared firstRehearse guest stop with storage online; check Proxmox task logs
Everything stays off after mains returnUPS never cycled its output or BIOS restore setting incompatibleTest UPS return-to-power and host auto-start behaviour

A communication-loss alarm is not automatically equivalent to a loss-of-mains event. Some NUT policies deliberately fail safe under certain missing-data conditions, while others depend on recent UPS status. Review the DEADTIME and notification behaviour in your installed NUT version rather than disabling warnings to make the logs quiet.

Security: Treat UPS Shutdown Commands as Administrative Access

A remote client with permission to force shutdown can take down every system on the UPS. Keep the NUT server on an internal management network, use the minimum necessary roles, limit TCP 3493 in the firewall and use strong, distinct credentials. Never expose the service directly to public IP addresses.

For segmented home networks, allow just the clients that need power status to reach the server. Keep the switch carrying the management VLAN on backed-up power. Home Assistant on a separate IoT VLAN may need a limited firewall rule for NUT telemetry; that does not justify opening full management access from all IoT devices.

Monitor recurring Data stale and NOCOMM errors. These are evidence that the automatic shutdown path might not be dependable when the next real blackout occurs. Treat the UPS software and its host as part of the backup infrastructure, including updates and configuration backups.

Cost and Hardware Checklist for a Reliable NUT Installation

NUT itself is free software, but successful automatic shutdown depends on compatible electrical equipment and working networking. For many home setups, paying for a good UPS with verified USB status and a replaceable battery is more sensible than buying a larger nominal-capacity model with poor management integration.

  • UPS: adequate measured watt capacity and enough verified runtime at your actual load.
  • UPS USB cable / management card: correct manufacturer interface; do not assume a proprietary connector is ordinary USB.
  • Switch: battery-backed Ethernet path between NUT server and clients.
  • Router: optional for LAN-only NUT traffic but helpful for WAN alerts and remote recovery; choose power placement deliberately.
  • Client machines: supported OS shutdown integration, tested permissions and correctly configured BIOS restart policy.
  • Battery replacement: local availability and cost over several years.
  • Backups: verified NAS and server recovery copies; UPS software cannot replace a backup.

For models and running costs, compare APC vs CyberPower vs Eaton UPS, NAS and Home-Server Electricity Costs and a practical 3-2-1 backup plan.

Frequently Asked Questions

Does the Home Assistant NUT integration itself shut down my NAS?

No. It reads information from an existing NUT server and can expose supported controls, but vendor NAS shutdown and Linux upsmon are separate mechanisms. Build an explicitly supported shutdown route for every machine.

Can one USB UPS shut down several computers?

Yes, if the UPS is appropriately sized and one supported server shares its status to reliable clients over a backed-up network. You still must install and test a shutdown mechanism on each computer. Monitoring alone is not enough.

Does a second NAS need a second USB cable?

Not usually when the first NAS supports a UPS server/master and the second NAS supports the corresponding network UPS client/slave role. Confirm the exact combination. This shares status; it does not create more UPS runtime or battery sockets.

Can I connect a Synology UPS server to Proxmox?

Some configurations expose NUT-compatible data, but vendor-documented supported clients and login behaviour depend on DSM version. Confirm the actual protocol, credentials, permissions and shutdown signalling in a supervised test. For maximum control in a mixed environment, a dedicated, documented NUT server may be easier to maintain.

What is the difference between OL, OB, LB and FSD?

OL generally means utility/line power, OB means operating on battery, LB flags low battery, and FSD is the NUT forced shutdown condition. The data come from the UPS driver and coordinator; verify your device reports them correctly during testing.

What happens if the Ethernet switch turns off?

Network clients may stop receiving UPS data. Keep the local switch between NUT server and clients powered by the UPS. A battery-powered NAS with an unpowered switch is an avoidable single point of failure for network shutdown coordination.

Should I wait until the UPS battery is at 5%?

Not by default. Many UPS runtime estimates become less accurate under old batteries or changing loads. Base shutdown on a tested runtime margin that covers your slowest guest and storage process; percentage alone is a poor universal trigger.

Will my mini PC automatically start again?

Only if the UPS restores output and your firmware/OS start-up configuration behaves accordingly. Some mini PCs will stay off after a graceful shutdown because their AC never actually disappeared. Verify this as part of the recovery drill, not during a real outage.

Final Checklist: An Automatic Shutdown You Can Trust

A successful setup has five independently verified parts: UPS state detected, network state delivered, clients halt safely, UPS/hosts behave correctly when power returns, and NAS data remains recoverable. If any one of those fails, a dashboard showing battery percentage gives a misleading sense of safety.

Begin with the simplest supported option for your hardware. One NAS can use its native UPS settings; a mixed Proxmox and Linux home lab can benefit substantially from a documented NUT server with dedicated upsmon clients. Let Home Assistant improve visibility, not become the only line of defence against an uncontrolled power loss. Then test the complete outage and restart sequence at least after major firmware, UPS battery or topology changes.

Related esp32.co.uk/ Guides

Official Documentation and Further Reading

CONTINUE EXPLORING / NAS & HOME SERVERS

Read next

Practical guides related to this article.

Explore the topic →
HARDWARE DECISION GUIDEPlanning your next NAS upgrade?

Compare hardware, storage and software before buying.

Compare options →