Smart Home Devices That Report Power Outages From Away
If your home internet dies with the mains, a battery-backed reporter alone can't reach you. Reliable power-outage alerts while away depend on two survival chains that both have to hold: a battery-backed device that reports the event to your hub (Zooz ZAC38, ZSE50, or ZEN15) and an internet path to you that outlives the router — with Home Assistant detection patterns for the outage and restore moments.
Last updated
The ugly version of a smart-home power outage does not start with a clean push notification. It starts with silence. The router drops. Cameras disappear. Cloud integrations time out. Home Assistant may still be running on a UPS, or it may not. A dashboard that looked beautifully green an hour ago now tells you almost nothing from 400 miles away.
That is why planning for smart home devices that help during a power outage means judging two separate chains, not one. First, something inside the house has to know that mains power failed and report that event to the hub. Second, the hub or security system has to get the message out to you over a path that still works after the router and modem lose power.
A battery-backed reporter can satisfy the first chain. It cannot magically satisfy the second. If the hub is dead, it cannot hear the device. If the internet path is dead, it cannot notify your phone. That distinction is the difference between “the house detected the outage” and “I learned about it in time to call someone.”

The two chains an away-from-home outage alert has to survive
When every mains-powered device stops talking at once, the tempting shortcut is to treat “offline” as the outage signal. Sometimes that is useful corroboration. It is a poor primary signal. A camera can go offline because its power is out, because Wi-Fi is down, because a switch crashed, because the camera rebooted, or because a cloud service hiccuped. A pile of offline devices may be emotionally convincing, but it is not a clean event.
The cleaner design is a reporter that sees mains loss directly, keeps enough battery to send the event, and is paired to a hub that remains powered long enough to receive it. From there, a surviving outbound path has to carry the notification. For a Home Assistant home, that usually means thinking about the hub, the Z-Wave or Zigbee radio, the network gear, the modem or gateway, and the notification route as one chain.
| Question | What has to be true | What fails if it is false |
|---|---|---|
| Who knows mains power failed? | A battery-backed device reports power disconnected/restored, or a monitored circuit reports abnormal voltage/power data before communications die. | You infer the outage from missing devices instead of receiving a clean event. |
| Who hears the event? | The hub and its radio stay powered long enough to receive and process the message. | The reporter may work, but the automation platform never sees the event. |
| Who tells you? | The home has outbound connectivity that survives the router/modem power loss, commonly through cellular backup. | The house may know the power is out while you do not. |
| Who handles restoration? | A separate restore automation confirms mains is back and reasserts the devices or states that matter. | Lights, plugs, cameras, pumps, and automations may come back in the wrong order or wrong state. |
The hardware choice matters, but it sits inside that larger survival chain. A plug-in outage reporter in a wall outlet is helpful. A plug-in outage reporter feeding an unpowered hub through a dead router is theater.
Start with a surviving signal source
A good outage signal is boring in the best way: mains was present, mains disappeared, mains returned. You can build around that. It is less ambiguous than “five things are offline,” and it gives restore logic a real event instead of a guess.
Zooz ZAC38: the clean mains-loss reporter

The Zooz ZAC38 is the tidy example because its outage role is explicit. Zooz describes it as a mains-powered Z-Wave range extender with an internal backup battery that can report “mains power disconnected” and “mains power restored” events to the hub. Zooz also specifies 24–96 hours of battery life from a full charge depending on network traffic, and says the battery is rated for 500+ full charges. Those are manufacturer specifications, not independent NestGrid test results.[1]
Placed on a circuit you care about, the ZAC38 gives automation logic a sharp event: power failed here, power returned here. That is much easier to reason about than watching a refrigerator plug, a hallway camera, and a Wi-Fi access point disappear and trying to decide which one represents the house.
There are two caveats worth treating as design constraints, not fine print. First, the ZAC38 still has to reach a powered Z-Wave controller or hub. Second, its event still has to leave the home if you are away. It is a strong signal source; it is not a complete remote notification system by itself.
Zooz ZSE50: outage signal plus local noise

The Zooz ZSE50 siren and chime belongs in a different slot. Zooz says it reports mains power loss and restoration, and can play a custom alert when it switches to battery. That makes it useful when someone may still be in or near the house: the alert is not only an automation event, it can also be an audible local cue.[1]
For an away-from-home owner, the ZSE50’s local alert is secondary. The remote value is still the same chain: the device detects and reports, the hub receives, and a surviving network path sends the notification. If that final path is dead, a siren may be yelling in the hallway while your phone remains quiet.
Zooz ZEN15: circuit and appliance clues, not the same job as a battery reporter

The Zooz ZEN15 800LR is useful for a narrower, appliance-centered reason. The Smartest House lists wattage, amperage, voltage, and kWh reporting for the heavy-duty power switch, which gives Home Assistant or another hub data that can help distinguish normal load changes from circuit trouble.[2]
That makes sense for a freezer, sump pump, dehumidifier, or other load where power behavior matters after restoration too. If the circuit is live and the appliance is not drawing expected power, that is a different problem from a whole-house outage. If voltage disappears before the device loses communication, it may provide a circuit-level clue. But a mains-powered switch is not the same category as a battery-backed outage reporter. Once its power is gone, its ability to keep talking depends on what, if anything, is still powered around it.
A practical design often uses both kinds of information: a ZAC38 or ZSE50 for the clean outage/restoration event, and monitored plugs or switches for appliance-specific follow-up when power comes back. That follow-up matters if the freezer, pump, or network rack is the reason you cared about the outage in the first place.
Keep the hub chain alive long enough to matter
The reporter-to-hub chain is the part that gets skipped in too many outage setups. A Z-Wave or Zigbee device cannot report to a controller that is off. Home Assistant cannot run an automation if the box hosting it is down. A beautiful outage event in the device’s firmware does not help if the radio stick, hub, or automation engine lost power at the same instant.
At minimum, the hub side of the chain needs enough backup power to receive the event, evaluate the automation, and hand the notification to the outbound network path. In a Home Assistant installation, that may mean a UPS for the mini PC, single-board computer, NAS, Z-Wave coordinator, network switch, and whatever else sits between the radio and the automation engine. The related question of sizing that backup belongs with UPS planning; see smart-home preparedness and UPS sizing if the hub chain is not already on battery.
This is also where protocol expectations need to stay realistic. Local protocols can keep talking locally when the internet is down, but local is not the same as reachable from away. For a broader compatibility view of what tends to survive offline, see which smart-home devices work during an outage.
The cellular-backup decision is the remote-alert decision
If the router, modem, fiber terminal, cable gateway, or upstream service dies with the power, your hub may still know exactly what happened and still have no way to tell you. This is the part automation logic cannot repair. Retries do not help if there is no path out.
Cellular backup is the usual answer for away-from-home notification. CNET frames cellular backup for security systems and standalone cameras as a way to keep alerts and remote control available when both power and home internet are down, typically at added subscription cost; its discussion uses roughly $10 per month as a tier example, but actual tiers vary by product and provider.[3]
There are different ways to implement that path. A security system may have its own cellular communicator. A camera may have a cellular plan. A router may support LTE or 5G failover. The important design question is not whether the home has “internet” in ordinary conditions. It is whether the notification path still exists after the same outage that removed mains power from the network rack.
- If your outage alert depends on Home Assistant cloud notification, confirm the Home Assistant host, network switch, gateway, and failover path all remain powered.
- If your alert depends on a security system, confirm the panel has battery backup and cellular service for alarm and status reporting.
- If your alert depends on cameras, treat cellular models or a backed-up router path separately from ordinary Wi-Fi cameras on dead household internet.
- If your plan is “the automation will notify me later,” decide whether a delayed notification is actually useful for your freezer, sump pump, pipes, or security needs.
The cost context is not abstract. Oak Ridge National Laboratory reported that major U.S. power outages, defined in the analysis as events affecting at least 30,000 customers or 15% of a county, averaged $67 billion per year in costs over 2018–2024 and reached $121 billion in 2024.[4] That does not mean your freezer will fail in any given outage. It does explain why a few minutes of notification-chain testing is a better use of time than admiring a device list.
Home Assistant patterns worth adapting, with their labels left on
Home Assistant users have developed practical outage-detection patterns around UPS state, voltage sensors, camera availability, and restart behavior. They are useful ideas. They should not be treated as confirmed behavior for your home until you test them on your own hardware. The patterns below are community-reported and unverified here.
UPS/NUT: “the hub chain is now on battery”
A UPS connected through NUT can expose whether the UPS is on line power or battery power. That is not the same as a whole-house outage detector. It tells you the equipment plugged into that UPS is no longer receiving normal input power. In the right layout, that may be exactly what you need: the hub rack is on battery, and the clock is now running.
The useful automation here is often not the first outage alert but the escalation. For example, a hypothetical Home Assistant setup might send one notification when the ZAC38 reports mains disconnected, then a second notification if the UPS battery falls below a chosen threshold. The first message says the house lost power. The second says your automation and network runway is shrinking.
Smart-plug voltage thresholds: circuit-level clues
Voltage reporting from a smart plug or switch can help identify whether a particular circuit is behaving normally. That is especially useful around appliances, pumps, and loads that should resume after power returns. It is less clean as a primary outage signal because the device may stop communicating as the circuit drops.
Use this kind of pattern for detail, not for the only proof. A voltage drop near the freezer circuit, followed by a ZAC38 mains-disconnected event, is stronger than either a vague “device unavailable” state or an isolated wattage change. After restoration, voltage and power readings can also help catch the more annoying problem: the house has power again, but the appliance you care about did not resume correctly.
Camera-down triggers: weak corroboration
A camera going unavailable is a useful clue only after you stop asking it to carry the case alone. Wi-Fi cameras depend on camera power, access point power, network gear, sometimes cloud service, and sometimes a vendor app. Any one of those can break the signal.
Camera-down automations make more sense as corroboration or security triage. If the outage reporter fires and the exterior cameras drop, that tells you which visibility you lost. If a camera drops with no outage signal, the better assumption is “investigate the camera path,” not “the home lost mains power.”
Home Assistant restart: possible restore marker, not outage detection
A Home Assistant restart after an outage can be useful as a restoration clue. It is not a clean outage detector because it occurs after the system is already down or rebooting. By the time Home Assistant notices its own startup, the original notification window may have passed.
Where this pattern earns its keep is post-restore housekeeping. On startup, Home Assistant can check whether an outage flag was set, verify that critical devices are available, reassert smart-plug states, and notify you that recovery is incomplete. For more on bring-up order after a storm, see how to restore smart-home devices after a storm. If your plugs or switches forget their intended on/off state, the related issue is covered in smart devices after a power outage.
Separate outage and restore automations
Do not make one automation responsible for both panic and cleanup. The outage moment and the restoration moment have different jobs.
| Automation | Trigger | Useful actions |
|---|---|---|
| Outage detected | Battery-backed reporter sends mains disconnected, or a tested local signal confirms the hub chain is on battery. | Notify immediately, mark an outage flag, record time, check UPS battery, reduce nonessential load, and avoid commands that require dead devices to respond. |
| Outage continuing | UPS runtime or battery state crosses your chosen threshold. | Escalate to a neighbor, security contact, property manager, or yourself with clearer urgency. |
| Power restored | Battery-backed reporter sends mains restored, or Home Assistant restarts and confirms the prior outage flag. | Notify restore, wait for network and devices to settle, verify key devices, reassert critical plug states, and clear the outage flag only after checks pass. |
The restore automation should be deliberately slower than the outage alert. Power can return unevenly. Routers boot before cameras. Z-Wave devices may take time to reappear. Cloud integrations may report stale states. A short settling delay and a second verification pass are usually more useful than firing every recovery command at the first sign of life.
For appliance circuits, restoration is when monitored power data becomes valuable again. A freezer that has voltage but no expected compressor activity over time is a different alert from “utility power returned.” A sump pump that never resumes after the circuit comes back deserves its own check. The outage alert gets your attention; the restore checks tell you whether the house recovered cleanly.
A practical build pattern
For a Home Assistant-heavy home, the grounded version looks like this:
- Put a clean outage reporter on a representative circuit. A ZAC38 is the straightforward choice when the desired event is mains disconnected/restored. A ZSE50 makes sense where a local audible alert is also useful.
- Keep the hub, radio, and necessary network gear on backup power. If the reporter cannot reach the automation platform, the event dies inside the house.
- Choose the outbound path before trusting remote alerts. Cellular backup through a security system, camera, or network failover path is the part that decides whether your phone hears about the outage.
- Use appliance or circuit monitoring for the loads that matter after restoration. A ZEN15-style monitor belongs here when wattage, amperage, voltage, and energy data help you verify behavior.
- Write separate automations for outage, escalation, and restore. Test them by pulling controlled power to the reporter and by simulating network failure, not just by looking at healthy dashboards.
That last test is the uncomfortable one, and it is the one that exposes false confidence. Pulling the ZAC38 from the wall may prove the hub sees a mains-disconnected event. It does not prove the message reaches you if the router also lost power. Testing the router, modem, gateway, or failover path is how you find out whether the second chain exists.
A reliable away-from-home outage alert is not a single device purchase. It is a tested chain: battery-backed reporter, powered hub path, and outbound connectivity that survives the event. Buy or configure the reporter for the power signal. Decide on cellular or another surviving path for delivery. Otherwise the house may know the power is out while you still do not.
References
- How Can I Monitor for Power Outages Using My ZEN15 Power Switch — Zooz Support KB
- Zooz Z-Wave Plus Power Switch for Heavy Duty Appliances ZEN15 — The Smartest House
- Best Home Tech Buys to Help During Power Outage — CNET
- Analysis shows power outages cost US electricity customers billions — Oak Ridge National Laboratory
Known issues with this device / protocol
Spec-version history
For active regressions on this protocol, see Update Watch.
No linked Update Watch entries yet.
