Not sure it’s really fixed. I just set up smart things for the first time this week, but this morning my presence sensor (mobile phone) properly sensed me leaving the geofence area, and executed the Goodbye routine which then switched to away mode and reported turning off my lights. However a quick check on the things tab showed the lights were in fact still on so I manually had to turn them off remotely.
Is this the exact symptoms everyone is reporting? Sorry, but really only a 3 day Smart Things user.
Your experiences are exactly what most of us are experiencing regularly (unfortunately). The outage mentioned in the last couple of posts was related to a mass-outage that they actually reported that they addressed, not the continuous issues we have been experiencing; which is what you describe here. I think during that outage period there was some sort of bigger system outage that affected routines that they reported. The intermittent issues you describe are the same as the rest of us and seems to at least partially go back to the cloud infrastructure in one way or another not being able to handle the amount of I/O since most of the system is cloud based and not locally executed commands.
I would think a volume issue would just result in a large slowdown. This seems more like a code bug or authorization issue where one of the final steps in the process link between local triggers (leaving geofence area) and cloud execution (sending off command to specific devices) is failing to actually execute even though the notification has properly occurred. Sorry to see this is not reliable at all right now, but the powerful integration has so much potential if they can isolate the issue and fix.
Somewhere on this board recently (I looked for reference and cannot find it) was an infrastructure engineer (that does not work for ST) that described over his years of experiences the same type of command “loss” through large systems. I too can speak on this to some degree with my experience in IT infrastructure. One would think that commands in a queue would eventually just keep piling up, keep processing, and eventually or hopefully catch-up, creating a large lag or delay in commands. That’s not actually what happens as the queue is flushed after a certain amount of time (that’s based on the system’s programming). So what you find are some events that are delayed and then some that just never happen. Likely what is going on with the ones that never happen is they hit the queue, they are not processed in time, and the queue is flushed before they are executed. So on your end, it looks like something just didn’t work or didn’t happen when really the command was never processed through their cloud and sent back to your local hub for execution. That’s why you are starting to hear people advise to move your routines off of the :00, :15, :30, :45 times. Even support just advised me to move my routines off of the hard times and move them a few minutes in one or the other direction to see if this helps… that right there shows that they are trying to have you send the command (request) to the cloud at a time when the system is less busy.
All my routines have stopped working, even when triggered manually. The system reports them executing, but none of the lights/locks/switches trip. I can trip the devices manually through the app one-by-one with no problem, but routines have been useless for over a week. This applies to routines with mode changes and those without. Deleting and recreating the routines has no effect.
This seems to be an ongoing problem. The platform is virtually useless without routines. I hope they address this soon.
Hmm, experiencing similar issues here now as of this morning. The Notifications tab claims that a routine was performed, but didn’t actually turn off any of the lights or change the alarm system status in Smart Home Monitor.
Concerned that this issue seems to be spreading…
Everyone seems to be having issues with the routines now… seems to be since Sunday/Monday of this week when they announced the new update coming tomorrow. I have asked support if this update will address these issues however am not hearing back. My gut tells me until they expand their data center and add a slew of additional servers to load-balance these requests; I don’t think we will see much consistent improvement.
Yes. We’ve been having this problem for a while now, for routines (turning on/off lightbulbs) that are solely time-based. Some routines work 90-100% of the time, some work but occur exactly one hour early, and the “Good Night” routine usually - but not always - fails to run at all. But it works when manually tapped.
My goodnight routine is not triggering. It will work manually. But I have it set to happen at 8:45. Hasn’t worked for a week straight. I have to manually push the button each night. It is supposed to shut off all my lights, and set the alarm to Armed/stay.
+1 for things not triggering properly. At this point I’m seriously considering ditching ST. Your whole business model is built around automation, how is it that the automation part isn’t reliable? I’d honestly stop spending man hours integrating new devices and throw the money at getting things fixed.
It’s a complete mess today for some reason. My devices weren’t reporting correctly in the app, I had to reboot the hub, and even after all of that, everything was hit or miss for about an hour where even my local devices (that are registered locally and execute commands locally and not through the cloud) were acting crazy. Then things started to come back and for now things seem to be OK but this is just crazy. I had to edit/rebuild all of my routines so they weren’t showing the next scheduled event in the past… this programming bug is just crazy to me. Why don’t they just flush these lingering “next event” routines after 10min or something instead of just letting them sit there forever. Hell, I had routines that still would work occasionally when they were supposed to but they were showing as not having run in almost a week. I just don’t get it.
I fixed my routine failures by following a post I saw on ST community. My routines were never %100…!! I deleted the failing routines, then I added new routines, renamed the routines that were failing and made some small changes to the times. BAM…!!! All my routines are working. …even the one that unlocks the door, turns off the alarm and turns on the lights (presence sensor) which hardly ever worked . May not work for you. …but it worked for me. .!!!
Yeah, I’ve done all of that actually and wasted a great deal of time doing so however for the past 4 days or so now, things seem to be working exactly how they are supposed to. Fingers crossed.
I can’t get my Good Night! routine to run. All my other routines seem to be running fine. 10pm rolls around the Notifications tab does not show it ran. I have to manually start it each night.
Yes, one of my routines failed last night too… so when this happens go into that routine and click through it as far as you can as if you are going to change the time of the event for instance. Once you get all of the way through, click “done” on each screen to back your way completely back out to the main routine screen. This “resets” that routine and will also reset the “next run time” to be scheduled when it’s actually supposed to run next. A lot of times when something fails to run, the system doesn’t recognize that and move on… instead it still sits there waiting for that time to execute again; so what that means is if your routine fails yesterday for example, ST is sitting there waiting for yesterday again and won’t trigger again until that is reset. The way you reset that is by doing what I describe above.
You can see your “next run time” for each programmed event by logging into the IDE then clicking on my locations, list smart apps, then you can click on each event to see the details which lists this information. Hope this helps.