Some of us have been trying to figure out how to approach an FAQ on what works during outages, and it just isn’t possible because there are too many tiny details that matter.
So if you need help coming up with a plan for dealing with outages, just start your own thread under projects and describe your exact setup and there will be lots of people who will be glad to help you brainstorm a solution for your specific needs.
Make sure you specify which hub you have, because there are some important differences. In particular, the new ADT/SmartThings security panel has a number of features that the other SmartThings models do not with regard to dealing with outages.
Also, it helps if you can indicate what kind of budget you willing to spend on your “Plan B” as that can make a big difference also. But you can wait on that one until the specific question comes up in your thread. Just be aware that some of the Plan B options may require buying additional devices or even services.
Meanwhile, here’s some background on what an “outage” means in the context of SmartThings.
For information on any current outages, see the official status page:
You can also subscribe to email updates on that page.
The 5 Points of Vulnerability
The SmartThings platform is primarily a cloud – based system. So that gives it 5 points of vulnerability:
1. Mains Power
2. The Hub Itself (device failure)
3. Local WiFi/LAN (The Internet might be up, but your own Wi-Fi router is not working)
4. Internet Connection
5. The SmartThings Cloud
When we talk about “local processing” in a SmartThings context, we mean what will run if the power is on, the hub itself is working, and the local LAN is working, but either the Internet is down or the SmartThings cloud is unavailable.
If you haven’t researched this yet, you will probably find that a lot less runs locally then you might think.
For example, the mobile app on your phone will not be able to talk to the hub unless the hub has access to the SmartThings cloud. This is true even if your phone and the hub are on the same local Wi-Fi. That’s not a technical requirement, it’s just the way the system is designed.
In addition, a number of things that you might think are handled locally are actually handled in the cloud. For example, you will not be able to change modes, run routines, or arm/disarm Smart home monitor unless your hub has access to the cloud. Again, it’s just the way the system is designed. ( and again, the ADT/SmartThings security panel is different from The other smart things hubs in this regard.)
No custom code can run locally at the present time. That includes webcore.
In fact, unless you have the ADT/SmartThings security panel, the only rules that can run locally are the official smartlights feature (and then only if all of the devices included in the rule can run locally) and a few bits of smart home monitor.
(And even the ADT/SmartThings security panel doesn’t change what can run locally in terms of your home automation rules. It’s just that the rules on the security side that can notify the ADT monitoring center will also be able to run locally.)
See the official knowledgebase article for the specifics:
Planning for Outages
As you start planning for outages, you need to consider your own specific household.
You can assume that there will be several SmartThings cloud outages every year, typically for a few hours but sometimes for more than a day, just because there always have been. An outage may be caused by a problem in the platform, or just be an outage for planned maintenance.
(The SmartThings platform does not allow individual customers to delay or refuse updates, they get pushed out automatically. Although the planned updates usually only last for 15 minutes or so, they can last longer.)
And sometimes a cloud outage doesn’t mean things don’t work – – it means that things work unpredictably.
For example, in a recent outage some commands were dropped but some commands were processed but with a delay of about 20 minutes, which meant lights were going on and off unexpectedly, the garage door might open at the wrong time, etc.
For this reason, during cloud outages some people just unplug their hub and wait until the platform stabilizes. That removes the problem of any unexpected commands running.
If you do that, then your planning will involve making sure that you can still do what you want to even without any Hub at all. Again, issues and answers will vary based on the exact set up, but it’s just something that you want to think through.
Another common situation is that the Internet is out or the cloud is completely unavailable. You don’t get random or unexpected commands, but you are limited to what can “run locally.”
Why plan at all?
Some people don’t bother to plan at all, figuring they’ll just walk over and use the switch if things aren’t working. And that may be all that many households need, as long as you do you have a switch that still works without the hub.
But there are a few individual cases where you may want to consider for a “Plan B”.
Plan B: Stairway lighting
If you normally use motion sensor controlled lighting for stairways, you want to make sure there is a Plan B if your home automation system is not working properly. That might just be a light switch at both the top and bottom of the stairs. But you will want something so that people don’t have to navigate stairs in the dark.
Plan B: Smart Door locks
Smart locks that have keypads typically store the user codes in the lock itself, so the keypad will still work as long as the lock itself still has power. But if your household is used to having the lock unlock automatically as they approach the door, you just want to make sure that everybody still has a way to get in if the home automation system is not working.
Plan B: Switches for Smart Bulbs
Most smart bulbs will turn onto full power when power is restored. Some people like that, some people hate that, but it’s just something to be aware of. (The Hue brand bulbs when connected to a hue bridge Will let you define the power restore action individually by bulb, which is really nice if, for example, you don’t want to wake up the baby. But most brands don’t offer this option yet.)
All smart bulbs will act like dumb bulbs if the home automation system is not available in that they will turn on when power is sent to them and they will turn off when the power is cut.
When we set up a home automation system, we often spend a lot of time coming up with some method of switches which can control the smart bulbs without actually cutting the current to them. That’s a good thing, because it means the bulb is always available to hear the next network command, and it protects the radio in the bulb from damage.
But it does mean we sometimes paint ourselves into a corner as far as outage planning, because we may end up with a switch which doesn’t work to control the smart bulb unless the home automation system is working fully.
There are a lot of different ways to deal with that depending on the exact model of bulb and switch chosen. But it’s just something to be aware of and to think through. If only local processing is available, how will you turn your smart bulbs on and off? There are lots of different possible answers, but which ones will be available to you will depend on your exact set up.
Plan B: Disarming the security system
Again, the ADT/Smartthings security panel works differently in this regard, and has more local processing options.
But it’s important to note for all other SmartThings systems that there is no way to change the security system status if the SmartThings cloud or the Internet is unavailable. If it was armed at the time the outage occurred, it stays armed until everything is restored.
Also if the system thinks you are away at the time the outage occurs, you will have no way to tell it that you have come home until things are fixed.
You need to take this into account when planning things like siren location. Many people are tempted to put sirens in some hidden or inaccessible location in the home to keep a burglar from smashing them. But if a siren is triggered during an outage, you may have no way to turn it off except to physically access the siren.
There has been more than one report in the forums from people who came home expecting Geopresence to disarm their security system, it didn’t because of an outage, and then it took them 15 or 20 minutes to be able to get to the Siren and cut its power before it would turn off.
So just make sure that your outage planning includes being able to turn off/disable any devices that you want to even if your system isn’t working.
Plan B: Smart Thermostat
Almost all smart thermostats have their own internal programming which will continue to run even if the hub is completely unavailable. However, if there is a partial cloud outage which is resulting in commands being run late or other flakiness, that can cause a problem when incorrect commands are received.
So planning thermostat management during an outage again depends on the exact details of your setup. But it is just something that you want to give a little thought to as having a thermostat that keeps resetting itself because it thinks you are away when you are in fact home can be a real problem for many households.
Most smart thermostats will run fine if the cloud is completely unavailable, although you might have to either walk over to the thermostat or use the thermostat’s own app to change some settings.
It’s the partial outages where flaky commands are received that can cause the most problems.
Thermostats and security features are probably the most common reasons why people will unplug the hub during a partial outage. It means you typically lose your light automations, but you protect yourself against unwanted commands for the other parts of your system.
A little planning can go along way not just towards preserving some of your automations during various outages, but also towards protecting your household from unwanted outcomes like sirens going off or thermostats dropping temperature.
And there are many different methods that can be used, depending on the specific devices you have and the specific outcomes you want to get.
One obvious example is just that if you want security features to keep running during an outage, the new ADT/SmartThings security panel offers many more features of this type than the standard SmartThings hub.
But even much simpler things like you want to make sure that the motion sensor for your attic light always works even if your Internet is out are situations that you can plan for, but which may affect device and rule selection.
So again, do feel free to start your own thread under projects on outage planning for your own household and people will be glad to help you come up with something which will work for you.
There’s no one right answer, but there should be some answer which will work for your specific situation, hopefully within your budget.
How often do outages occur?
SmartThings pushes out firmware updates every month or two. We typically get three or four days notice ahead of time, but not always, and these updates cannot be delayed or refused. Usually your hub will be out for about 15 minutes, but sometimes it’s hours. When the hub returns, everything might work just fine, or you might have to reset some individual devices or reopen the app and save some rules again. And every time there’s an update some small but significant percentage of users end up with problems afterwards, and you never know if you’ll be one of the unlucky people.
These planned outages are not posted to the official status page, but as customers, we have to take them into account, since they still take our systems off-line.
In 2017, there were 7 planned outtages, but some of them had problems and resulted in multiple outages over several days.
In addition, there have been several unplanned problems every year. Sometimes these only affect one region, sometimes they only affect one device. And sometimes it takes weeks before those are fully resolved, such as a recent problem in the UK region where modes were being ignored. There were at least 12 of these unplanned outages which were not device – specific in 2017. Most were resolved in a few hours, but not all. These were in addition to the planned outages.
The first bug reports page in the community – created wiki gives a pretty good picture of what kinds of problems occur and you can read the community discussions around any individual ones.
The company has repeatedly stated since the spring of 2016 that improving reliability would be a top priority, but clearly outages do still occur from time to time so it’s best to have a plan for what to do when they do happen.