Stuck Sensors and Rate Limits?


#1

I noticed several different people reporting issues with sensors being “stuck” over the last week or so. That is the sensor appears to say that it is always in motion, or always open, or something like that but the person can see it’s not being triggered. And it may be several devices all failing at once.

Physically, it’s very unlikely for this to happen. it would more likely be a failure of the status reporting mechanism, in this case the smart things cloud or the smart things mobile app. So that status was not being properly updated.

It occurred to me that this might be related to the new rate limitations that smartthings introduced in the last week. These are designed to limit any individual smartapp from firing more than 250 times per minute.

If you are issuing polling or refresh requests for a whole bunch of devices with a very quick interval, you might hit this limit.

It’s not 100% clear yet what behavior you would see if you did hit the limit, but it’s not unreasonable to guess that it might be something like a sensor status not getting updated.

So if you do have “stuck” sensors the first thing I would do is look at the activity logs and see how much activity there is, and how likely it is that it’s coming from the same smartapp.

More information on the new rate limits here:


(Bruce) #2

A properly functioning motion sensor can’t come anywhere near to throwing that many events. Once they sense motion, they don’t transition back to inactive for many seconds after the motion ends. The only thing that could hit the rate limit vis a vis motion sensors is a defective one, or a defective app that goes into some kind of loop.


#3

Agreed, of course. But the people who were reporting stuck sensors were reporting multiple sensors all failing at the same time.

The rate limitations are per smartapp, not per device,

So my thinking was that if they had refresh going on a group of devices, then they might be hitting the limit just because of the poll requests, whether the devices were asleep, awake, or ignoring the requests.

The new rate limitations might not have anything to do with it, of course. But they are a new back end change during the same time that some people are seeing some really weird status reports in the mobile app.

Might just be a problem with the mobile app itself, given that it also changed over the same period. But then I would expect more people to be seeing the problem with more kinds of devices.

Normally if there’s a status problem with a bunch of devices at once, you look first at the routing, and second at the hub antenna. But neither of those seem to fit what was happening, particularly since it was happening in multiple people’s houses. But not everyone’s.

Who knows? It could all be a coincidence. But I just thought I’d mention it as a possibility.