I am curious how many here have had some of there apps just go AWOL. I have an app that monitors temps on equipment on a low voltage board in my garage, and turns on fans when it’s warm.
Well its 93 here and the bloody thing hasn’t been triggered all day, despite seeing high temps reported in the sensors activity log for hours!
So I pop it into the simulator and bingo, the next temp event triggers the apps handler.
I could have sworn I’ve seen this before (but wasn’t sure); now I know. I’m ok with having to be married to the cloud for everything so longs as it works, but far too often it seems that it does not!
LOL, dealing with the same damn thing. The event subscriptions just don’t fire reliably. So I set up a polling for sensors and posting them to thingspeak.com
I figure a schedule to grab latest device values every x minutes is the best way to go for now.
But have you had just plain stop until a sim load or a reinstall?
This one was running fine for a week as I had it sending me push notifications. Then, yesterday I turn them off and it is broken all day today. Dozens of target temp events; not a single trigger.
The forums are an awesome place to collaborate but for standing issues they generally get fixed faster if you report them to the support team.
I am also having this issue. My biggest thing is the trigger response times. Sometimes it is instant and sometimes it can take 3-5 seconds. I know they are working hard on this aspect of the cloud. And last i heard they had things down to like 300ms round trip. Device up to the cloud then back to the device. Also, last i heard the entire company is focused on reliability.
ST has their priorities in the right field at this point
Did you mean “in the right field” or “in right field” LOL. But seriously, I came here not to resolve it, but to see if it was common behavior. If it breaks again, I will report it while it is broken.
But it is possible that it was my lazy code that broke it completely… I had an ignore state set when the fan turns off, that I reset X minutes later in a schedule handler. If that schedule handler doesn’t fire, my event handler is dead. Maybe this kind of logic puts too much faith in SmartThings being able to do what it is told.
I think for important tasks, we’re going to need some error handling. For instance, how can my app know the schedule handler won’t execute… that either the SmartThings’ servers didn’t receive it, or after being received, it somehow fell out of the queue? Is there some form of response that can be parsed to know it was received, or a function I can call periodically to confirm my scheduled events are still pending execution? and is there an event (other than poll) that works reliably (my experience so far has been that poll does not).
Now I have only been at this a couple of months, and these kinds of things may “be in there” someplace, but every App example I have looked at just expects things to work, with no error handling code of any kind. So if they don’t, we simply stand by while our equipment melts down or our basement floods; unless we are present and have the wherewithal to notice our automated devices are no longer automated.
Since this stuff is all running in the cloud, we need for the could to police itself. But more importantly… we need the cloud to communicate with the SmartThings app when an error is encountered, with an intelligible response in as timely a manner as possible (and not through the hub, but directly). Perhaps some kind of realtime audit process tracking events, and verifying their execution. I realize this would be redundant as I am sure the server side stuff has some level of error checking/handling, but at this point it doesn’t seem that it is enough.
You can build some of this into your app. For isntance an app that i have to turn on my porch light turns it on then checks the last status to see if it is actually on then if its not it sends me a notification.
Im sure this stuff is natively coming at some point to STs but in the meantime you have to buid that logic in yourself.
This is the beauty of buying into an open system If you see a down fall, report it then fix it yourself until an official solution exists.
I am also having this issue. My biggest thing is the trigger response times. Sometimes it is instant and sometimes it can take 3-5 seconds. I know they are working hard on this aspect of the cloud. And last i heard they had things down to like 300ms round trip. Device up to the cloud then back to the device. Also, last i heard the entire company is focused on reliability.