After reading this post, from the perspective of someone who has been on the fence about home automation and considering SmartThings for over a year now, my question is this: Is there a good summary from the community of what currently isn’t affected by instability/load/responsiveness issues on the platform side?
If basic manual on/off and timed scheduling for half a dozen SmartThings Outlets works reliably that would be enough for me to warrant going forward, but if I can’t get some set of basic stable functionality then the whole thing will cost me more time than it saves.
I’ve been watching the ongoing problems on a variety of threads since the unfortunately titled December 2014 “Smartthings: I’m sick of this sh*t! Shape up!” thread, and waiting for a general sense of “Ok things are better now” before I bother moving ahead.
I don’t mind waiting, but in addition to all the users who are frustrated with their current systems I’m probably not the only one who’s scratching their head as to whether the platform is worth getting into.
I think it has been mentioned elsewhere that each person’s experience is going to vary but in general I would say that if you are looking for the basic functionality that you mentioned I think your experience will be ok. You may have an issue with the occasional device falling off the network and needing reset but other than that I have not experienced any major issues that would make me tell someone to avoid the platform for that use (remote control and basic automation).
If you want to use it for the more advanced automation and have complex rules or if you want to use it as a security system at all my recommendation would be to hold out a bit longer.
These are only my current views and others likely have other opinions.
(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy)
My very simple sunset related timed light switches (4) failed last night, though they had not failed in quite a few days since the push of the new scheduler (“Ticker”). Before the scheduler upgrade, failures were more common than successes.
Statistically it’s hard to say what that represents. 1 day failed out of 10 … 90% reliability? I really need to be able to expect much much better than that, but everyone’s needs and tolerance differs.
Alex is expecting a lot of patience (and, frankly, “support”) from us, his Customers. I appreciate that he’s forthright and asking; but … same answer as above: Everyone’s patience level is different.
I agree, everyone’s tolerance for failure/unreliability is different, and from reading the forums there are still some number of users whose use cases seem to work reliably the vast majority of the time. I don’t mind things breaking every once in a while, (one failure every 10 days would be annoying, but usable as long as it eventually improved) and am definitely happy to troubleshoot and contribute.
I’ve worked on internet-scale products from startup to hundreds of thousands of users, so I can relate to exactly where the SmartThings team has been over the last couple of years. Selfishly, though, as an end user, that doesn’t remove some level of basic “If I’m going to bother spending $400 and more importantly the time to set this up, I at least want my lights to turn on when I tell them to most of the time, or I might as well just keep flipping switches myself.”
I’m happy to stick to non-critical things like light and power outlet controls, where failures are annoying but non-destructive for quite a while before going any further, and even if light control is all I got then it’d be a benefit. Of course given SmartThings’ support for multiple automation platforms and a huge range of devices, the goal would be to use it as a central point of automation for a wide range of things around the home, but again, I can wait on that and deal with some growing pains if I can get some short term benefit.
In spite of the long-running issues and complaints in the forums, the general sentiment that “We love this product but wish it worked more reliably” and general positivity of the community is what’s led me back to SmartThings vs. WeMo, Iris, Wink, Google/Nest (though time will tell on Weave/Brillo and I may decide to just go early adopter there).
Ultimately though it sounds like the platform instability has been affecting everything - which I’d expect. I’ll probably see what others say, do some more reading and check back in a couple weeks.
My setup has been pretty reliable. In the past occasionally it wouldn’t go into night mode or good morning Mode. Which isn’t that big of a deal because if it doesn’t I just do it manually. But now with the new schedule it’s pretty rock-solid. It sounds like you want to stick with the basic system. That’s not going to happen. This stuff is like a drug. You want more and more. And if you’re lucky you won’t drive your wife nuts with it. My wife has gotten used to my addiction. It was pretty rough early on. Just go ahead and do it. If you regret it you can come back and yell at me. And hell, I might even buy your stuff
Ditto, no improvement whatsoever. The status should not have been changed to Resolved. This is more infuriating to me than the issues I have been experiencing for the last month. ALL SYSTEMS OPERATIONAL my ass.
It sounds like you want to stick with the basic system. That’s not going to happen.
Yeah, I’ve resigned myself to that already, starting off basic is just a way to say “Ok even if all the nonessential bells and whistles muck up sometimes, at least I’ve got the basics that work reliably.” Believe me there are several projects beyond the house stuff I’d like to tie in down the line. You’re convincing me to go for it though.
Now that I have a better understanding of the platform they are running on it makes a great deal more sense on the failures I have seen over the past several months. It explains why some people experience problems, while at the same time, other don’t. It explains why the state of things are out of sync. It also seems to explain the why some rule machine rules go headless and others don’t
The question is, is the problem fundamental to the platform? And is there a way to fix it without a total redesign?
It’s not that they should or should not be using Cassandra. There are many very good reasons why Cassandra can be the best, and in some cases, only suitable database for high scalability situations. However, when things are going badly with Cassandra, it usually ends up looking like this to the end user. There’s a very long list of things that can go wrong. At the basic level, with Cassandra, your data model can directly affect the reliability of the database in ways that people new to Cassandra would not expect. For example, too many reads and writes to the same row can create “hotspot” situations where the database simply crashes. Dealing with “eventually consistency”, where you can tune the consistency settings for a given read or write, can have unintended consequences as well. Depending on your consistency settings, the performance can slow down quite a bit. These are just the basics, and a lot of challenges with Cassandra only turn up as the scale goes up, so it’s very common for people to build systems that work beautifully and then abruptly hit a point where nothing works anymore. At that point, they initially attempt to solve it from an ops standpoint, adding more nodes to the cluster, etc. Eventually they realize the problems stem from application-tier choices and they have to change all their plans. Not saying that’s happening here, but it fits the profile.
I’m no Cassandra expert, so I’m not going to pretend to give some answer I can’t guarantee is right.
What I will say, is we have brought in some new engineering fiolks and one I am particularly excited about comes from high frequency trading (so do I actually, but I supported the platform he traded on). So if anyone knows how to do DB read writes at an insanely high rate, while remaining accurate and reliable, this guy does.
He can’t change things overnight of course, but with his guidance and some other new engineering folks we’ve brought on, I really do believe (for real not just saying this for fluff) that we will begin to see incremental and meaningful improvements.
We are willing to earn back your trust, and honestly, that’s the only way I’d want it, we can only humbly ask for patience as we grow and work harder everyday to earn your trust back.