Notify me when a device goes offline?

I plan on getting a battery backup on my internet. I have a ST outlet and a GE outlet. Id like to be notified when power is lost to these devices, indicating that power has been lost to the house.

Main problem here ,is that if there is no power to the plug, than it has no way to tell the hub that it has no power.
There was talk about a notification from ST if it was on battery power, BUT AFAIK it has not been implemented yet.

I suppose that would be a possible solution. I was hoping to have the hub on the same battery backup as the Internet (and surveillance cameras) that way I can get maximum run time from the hub.
Doesn’t the hub communicate with the devices on a regular basis? Why can’t I be notified when it attempts to communicate but doesn’t get a response?

Hub has it’s own battery backups, so not really any reason to use UPS. Yes the hub does poll devices, but I am not sure how the logic would work to have it report when a device didn’t report back.
A lot of devices only report if there is a change of state. In the case of a plug it sends a signal if you hit the manual override button to turn it off/on, or if it has a power meter it will report back at set interval if the usage has changed.
So I guess in theory if you had a power reporting plug somebody could write an app to send an alert if it didn’t report any usage for X time, but that would lead to a lot of false alerts .

The nature of mesh networks is that any given device may not respond to any given command at any given time. There is absolutely no guarantee that a message will make it to a device, and there is also no guarantee that a device will respond. This is all under normal circumstances, so the hub isn’t concerned at all if it doesn’t get responses back from its nodes.

However, there are some devices that will explicitly report their power status. The Gen 1 Smartthings motion sensor was one of these devices (you can still find it for sale online) and the MIMOlite is another. The MIMOlite does this by using energy stored in a capacitor to send a signal out when it loses power. The Gen 1 Motion sensor does this by reporting when it switches from USB to battery power. There are smart apps that were specifically written for these devices.

Your GE outlet and your ST outlet don’t report their power state, and your hub can’t rely on getting a “response” as an indicator of power.

Hope that helps…


Does the Hub not send out heartbeat monitoring of its connected devices?
I’m still pretty new to this but the App reports the status of devices and it also logs when my connected devices are unavailable and when they become available.

SmartThings is offering paid monitoring of Security systems. I’m wondering if there is an app or how difficult it would to create one such that when the Hub detects a device is unavailable it sends a push message to my smartphone, or if the device has been unavailable after a preset time to send a push message.

The hub isn’t designed to periodically ask if it’s devices are alive. That would be a burden on the mesh, and it would needlessly drain the batteries of battery powered sensors. Generally, sensors are designed to send messages to the hub only when something about their state changes. This is a reality of most mesh networks. It isn’t unique to ST.

How about get a cheap internet switch and plug it into GE outlet. This switch is your ST hub internet connection. Power out on plug = no internet to hub = notify hub offline. Sounds strange but it’s an option.

I stumbled upon this thread after one of my door sensors failed to report an open door.
The hub said everything was fine (Connection good, door closed, temp fixed, etc.) When I replaced the battery on the sensor everything was fine again.

What I can’t wrap my head around is people relying on this technology to automate tasks or as a security system. If the system can’t report lost or failing sensors and assumes that no news is good news then what’s the point?

That’s the nature of mesh networking. It’s my understanding that proprietary security systems that use mesh networking apply some additional logic to the system to attempt to correct for this. However, this means that both the sensors and the hub need to be speaking the same proprietary language. @JDRoberts might have more insight into this issue.

SmartThings is not a reliable security solution. That’s just the reality.

1 Like

There are several different types of network topologies. Wi-Fi and hardwired both use either a star or tree topology where there is continuous feedback between each device and the network coordinator, and it’s really easy to know as soon as the device goes off-line. In these type of topologies, there are specific message routing paths that are defined by the structure of the network.

Zwave and zigbee Mesh are mesh networks. Mesh networks were originally designed for massive sensor nets where you might have 500 or 1000 battery operated devices and you really didn’t want to have to create a whole new routing table just because one device was off-line getting batteries changed. So mesh networks do two things: first, the Coordinator doesn’t really care if any individual device is unavailable, and second, the routing paths are fluid and devices are allowed to try alternate paths if a particular neighbor is not available. Again the whole point is not to have to worry about any particular device being pulled offline.

making mesh act more like a star network

Obviously, as has been mentioned, basic mesh doesn’t really meet the needs of most security networks. There are multiple ways to address that.

  1. Force the devices to use a known path, usually by keeping them within one hop of the Coordinator. That works in a lot of residential situations just because the area covered is small enough.

  2. perform a “wellness check” at regular intervals. This is where you ask a device to respond and then take action if it doesn’t. It’s true that this is sort of the opposite of the original mesh philosophy, but some security systems do do this. Iris, for example, will let you know when a device fails the wellness check, but they only do it once a day when they check the battery level. It’s not the same thing as being notified immediately when a device goes off-line, but it can meet the needs of some use cases.

The obvious next question is how often should you do a wellness check? They do take up system resources. Some systems will do one once a day, some once an hour, some will do them once every 15 minutes, it’s rare to have them done more often than that because then you would probably go to a different type of system instead. But it’s not impossible. The fewer devices, the easier it is to spend system resources on wellness checks without impacting other activity.

Three) use a tamper alert on the device itself. This is usually an accelerometer, although there are some other methods used as well. A lot of door locks have these. The idea is that if someone intentionally tries to take the device off-line, or even just physically move it, it will send out a panic signal.

So these are all methods that are used to take advantage of the flexible and inexpensive nature of most mesh devices while still including them in a low cost home security system.

SmartThings doesn’t have a built-in wellness check option the way some other systems do, but you could certainly write one pretty easily. The forced pathway Issue is just a matter of how you’ve laid out your network. And you can buy devices that have tamper alerts if you want those.

But there’s a reason why almost all zwave device manufacturers include a warning on their user guides that the devices should not be used as part of a medical support system. And some of that same logic applies to most critical security systems.

But obviously it’s a personal choice. Everyone has different requirements for peace of mind.

I personally use a completely separate system for security, one that is UL listed, but it has a monthly fee, and a lot of people don’t like that.

So different things work for different people. But if you want to use a mesh network for home security, I’d definitely consider adding some of the options I’ve mentioned.

BTW, there’s been quite a bit in the tech news lately about using jamming signals to prevent security system sensors from going off when motion or movement is present. A wellness check catches these because the device will not report back when queried. But again if you want one that runs more often than once every 15 minutes, mesh may not be the right match for that particular use case.

1 Like

I do too. And I don’t like the monthly fee, but I still pay it!

1 Like