ST has become more unreliable over last few months


(Coolcatiger) #1

I am on ST platform from day 1.

I still feel like I am beta testing ST. It works many times but over all performance is not of production grade system.

Even though ST support is quick to reply I have stopped contacting them just because I am not always sure if they will fix issues forever.

It looks very polished product but it’s not reliable. It has even damaged my garage door and car costing me $$$. After that I stopped using Automated Garage Door app completely.

The story of reliability of presence tags is another topic.

So far i have invested few grands in ST platform. It’s much better than X10 but still not what I hope for.

What do you think guys ?


(DanG) #2

I agree. SmartThings has enormous potential but the flakiness continues to weigh against it. Like you I have invested well over $1,000 and I sit and hope they get the reliability to 99.9. My main interest is security and for the most part it works, but… Having a reliable connection to the internet is one issue, one that I have control over but their back servers are only in there control and it seems that is where almost all the issues are. I am still waiting for the Android app to be redone so it makes logical sense. Half the time I can’t find where anything is because it is so chopped up.


(Convinced ST will never be unbroken…) #3

Yeah… it works, but it fouls up often enough for me to consider it unreliable. Therefore I don’t use it for any mission critical or security related tasks.


(Beckwith) #4

Can you expand on this so others can avoid this?


(Ron S) #5

I am new to ST with just 20-25 things. Like everything about ST but the reliability factor. I have avoided using it for critical things as well such as my garage door, front door lock etc. I sincerely hope them the best but the servers has to respond as expected each and every time else it is a very serious issue.

I went thru their status page and every month there has been “cases”. I hope ST is listening and improve on this.


(Coolcatiger) #6

@beckwith
I was using Garage Automation app to open and close garage door when presence sensor comes and leaves. One day actually the 2nd day I was using the app the shutter closed on my car’s trunk because ST though my presence left, this damaged garage door and trunk of my car.


(Rob Klemm) #7

@coolcatiger - Didn’t your garage door auto-reverse when it hit your car? Perhaps the resistance necessary to reverse is set too high? It really shouldn’t damage the door if it gets stopped by something. Otherwise, a LOT of people with kids would be replacing doors CONSTANTLY! The idea of putting this system in charge of anything that can cause harm/damage like that is definitely a long way off for me. I actually have my whole house fan hooked up to a switch, but ANY time I go in the attic, I pull the override shutoff tab on the switch, for fear of the system going all HAL 9000 on me while I am up there.

As far as up-time goes, it is definitely quite concerning. My trepidation with going ahead with ST was the fact that my internet connection might not be completely reliable (new house with Comcast in the woods just seemed like a recipe for disaster), but thus far my connection has been rock solid. I never thought the issue would be with the back-end servers.

To the ST staff, fixing this really needs to be the top priority. If your earliest supporters are doubting the reliability, then the platform’s long term prospects are in trouble. Although, with the recent acquisition, maybe that is of no concern? I hope that is not the case, but I have seen first-hand what can happen when a huge company buys a little one.


(DLee) #8

While I share your frustration when ST instability happens, I don’t think you can bucket your garage door incident into this category. The limitations of presence sensors are well documented to the point that I never even considered using the rid automated garage door app. There are other things you can do to mitigate the risk to human life and vehicles from any sort of automatically closing garage door. Reducing the closing force of the door, or adding more IR saftey sensors to cover your non-standard case of having the trunk hatch in the line of door’s closing path. Or setting the presence delay to 5 or 10 minutes etc which is necessary to use that device in the first place, yet makes use of the automated garage door app unfeasible for most.


(Beckwith) #9

@coolcatiger

Thanks for letting us know what happened. I’ll make sure not to program in any automatic close scenarios.

I notice flashing light and audible warning signal is now required before it can get an UL listing. I’ve worked with robotics and you can never have enough safety measures with unattended execution.


(Chrisb) #10

FWIW:

I’ve found ST to be in the 90+% reliability area. Probably north of 95% actually. That’s still low enough that I’d never rely on it for something like managing a webserver that needs to be up 100% or for living saving equipment. But for the most part I’ve been pretty happy with it.

Only relatively recently have I started using it for high levels of security. I wouldn’t say it’s on par with monitor services… not by a long shot. But on the other hand, I wasn’t ever planning on getting any sort of monitor service so really it’s just kinda a bonus on top of my home automation.

RE: Presence sensors: Once I put a SmartSense motion sensor… plugged in… in my garage, my sensors have been rock solid. In the many months since I did this the only time I suffered from the wandering sensor issue I track down to be an problem with the sensor. Someone had bumped it enough that the power cord came out and it was just running on battery.

Generally I’ve been happy with ST. I certainly understand if others aren’t. I may just be lucky as well that I haven’t noticed the delays or downtimes as much (though I did notice issues last night for a while). But I also feel that anything you introduce complexity into a system, you have to expect additional problems. You know what will never fail? Putting the dumb switch back in the wall and manually flipping on and off your light.

I know that sounds a bit condescending but it isn’t meant too. I’m just trying to say that with a more advanced system we have to expect some level of problems. with it. Of course everyone has to decide what the acceptable level of problems they will tolerate before they throw the complex system out. For me, the current performance level is sufficient to keep me from looking to throw it away. Not that I don’t want better performance, but I’m not at the point that I’m ready to scrap anything.

This might be partly mitigated by the fact that there are perhaps only three things that “automated only” in my house. That is to say, nearly everything still have a manual override ability: Light switches or button on outlets can always be used to turn on/off things. I will have my normal garage door openers as backup for the automated system. With a non-geek wife I’m been sorta locked into make sure that she could still use the house in the event of a total ST failure. Nothing would destroy her acceptance of ST faster than her not being able to do something that she needs/wants to do because ST happens to be in a rare “unresponsive” mood.

Final note to a terrible long post (sorry): It’s worth nothing that the great majority of my things (and I think most people’s things) are NOT ST proprietorial. Most of my things are Z-wave. And I think even the ST zigbee stuff is standard Zigbee HA so the majority of my hardware should be movable to another platform. Obviously software side any work I’ve done is out the window… though I have learned quite a bit doing it which should translate… at least in part… to a new system.


(zraken) #11

You did what with the Garage Door ?!
Ok, it’s time for ST (and all Home Automation vendors) to put warning labels on their products. Even with a 100% reliable “System” there’s always going to be edge cases where things don’t work the way you envision, so you want to think twice before adding automation to stuff that can cause potential harm.
Having said all that, I think the ST System Design is still not usable even for simpler automation - for instance I’m having doubts about using it for sensing water level in my sump pit (to check for sump pump failure) and turn on a Siren in the house. What if there’s a couple of internet HD video streams going on at that time and the packets between the Hub and ST Server don’t make it through fine. So many things can go wrong in this path without the Internet completely going down.
As it stands, ST is a gimmicky cool thing you can show off to your friends (which I do a lot). Please don’t use it for anything critical and complain later.


(Kristopher Kubicki) #12

I actually had this happen to me too on a Vera. I opened the door via a momentary relay, but it didn’t open right away. I hit it again and the door came up. However, the delayed second command actually closed the door while my Cayenne was halfway in. The IR sensor somehow didn’t stop the door from coming down (I tested it like 50 times after… without issue) and it put a nice scratch in the roof. Fortunately the door went right back up right away.

I don’t really have anything of value to contribute to the thread, but that series of failures really bummed me out.


(Kristopher Kubicki) #13

I’ve found ST to be in the 90+% reliability area. Probably north of 95% actually. That’s still low enough that I’d never rely on it for something like managing a webserver that needs to be up 100% or for living saving equipment. But for the most part I’ve been pretty happy with it.

95% uptime means its not functional 1.5 days out of the month. :frowning:

My biggest complaint (aside from the last couple days, which has been totally awful) is the latency. I have a very low ping to the ST servers, but commands that should fire in <300ms sometimes take a lot longer. If the hub was doing some of these activities local, it would be no problem. I think with ST2 this is going to happen.


(Coolcatiger) #14

I can give lot of examples where it does not work when it is supposed to work.

e.g. i have multisense attached to the basement door and have an app programmed to turn on basement lights when i open the door. Some time it works but many times until i reach the to the bottom of stairs light does not turn on. Because of this my family just uses switch to operate basement light. But since I am burned by investing money in ST I am adamant of not using switches and always reach to bottom of dark basement and wait until ST shows it’s magic. And my family then has a good laugh :frowning: Atleast this is the incentive :wink:

I have about 70+ devices , but i don’t this is an excuse for delayed processing…


(Beckwith) #15

@zraken

If you are using a water sensor to tell the pump to go on and off, I agree, a manual float switch is better. However, SmartThings is great for notifying you when the pump fails. Before SmartThings, I had a battery operated water alarm. Of course the batteries went out and we were away when it issued its low battery chirps. Being able to monitor battery and get push notifications is a huge improvement over previous options.

So for me the new capabilities outweigh the occasional downtime. I’m building out with the assumption that uptime will improve over time but make sure I don’t get to far ahead of the technology.


(fxstein) #16

Reliability is a major factor. This week alone my Hub became unresponsive several times and the Nest communication has completely stopped working.
Putting my HVACs/NESTs into Away mode based on presence stopped working 2 weeks ago.
Still baffled that I have to run Pollster to get any meaningful device updates for the rest that still works. Points to a lack of capacity in the AWS cloud or a non scaleable design - both not good.

Compared to that my ISY994 local controller for INSTEON (150+ devices) has been up without interruptions for almost 200 days. And that downtime was cause by electrical installation work resulting in shutting down power to the house.

I would also recommend to put some math behind above mentioned uptime claims:

http://uptime.is/99

95% 1h 12min downtime a day
98% 29min downtime a day
99% 14min downtime a day
99.5% 7min downtime a day

It takes 99.94% to have less than 1 min of downtime a day. That is still 5h 15min per year. Its worse than what I get from the local power company.

Still hanging in there, but stability or lack of is a major concern and there are a lot of new platforms coming out promising high availability or off cloud.


(Geko) #17

Power companies are public utilities. It seems a bit unreasonable to me to expect the same level of reliability from a free consumer-grade service. Not that I’m happy about outages, but you get what you pay for.


(Convinced ST will never be unbroken…) #18

Hmmm… Possibly a new marketing slogan for SmartThings? Or maybe “We don’t charge for our service, so don’t expect that’ll it will work all of the time.”

But seriously, there is no higher, paid level of service with SmartThings so we really can’t evaluate it as “you get what you pay for”. It is more like “you get what you get” with SmartThings; and that means your house doesn’t always do what you told it to.

BTW, came downstairs this morning to see all of my “middle of the night” lighting coming on. Seems my mode change from sleeping to home at sunrise did not fire today.


(Sarcjack) #19

I had the same problem with events not happening at sunrise. Second time this week. Is there any way to drug this?


(Geko) #20

Like it or not, but it’s true. Going from 95% to 99% reliability is relatively easy. Going from 99% to 99.9% is very hard. Beyond that - is next to impossible without fault tolerant components throughout your system and is very expensive.