tgauchat
(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy)
#431
Every system is going to have some degree of failure which is loosely proportionate to (cost + flexibility). i.e., a rigid system is less likely to have problems (Staples Connect is mentioned a lot); and expensive systems … well, the point is loosely proportionate.
If the concern is the classic "I’ve fallen and I can’t get up"tm, then get that Philips dedicated system for that purpose.
If there are a short list of critical items, then, perhaps, just perhaps, you can still use SmartThings as the heart of that system with a few bits of redundant hack code, etc…
Medical monitoring systems were recently discussed in the following topic. Although The use case that initiated that topic is about someone who fell and blacked out, all of the commercial solutions mentioned for that problem also have a manual button that can be pressed to initiate the alert on demand, so they also solve your problem.
1 Like
tgauchat
(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy)
#434
So why isn’t ST a “decent” option? For something as simple as a button that activates a light or bell in another room, what is SmartThings’s outage rate? If 95% uptime can be established for that function, what are the odds that she requires assistance at the time time there is an outage? i.e., do the math … requirement frequency x outage frequency > x is too big or not?
As someone who needs the ability to contact an aide in an emergency there is no way I’d rely on SmartThings as it currently operates.
First of all, 95% uptime is 5% chance of, well, dead. That wouldn’t pass any medical monitoring standard I know of.
Second, there’s that whole “Here’s an update that will take your system offline for an indeterminate amount of time with an undefined restart procedure afterwards” which happens 2 or 3 times a month. Again, not something that would pass a medical monitoring standard.
There are many alert button systems available that work great, cost very little, and have uptimes above 99.99%. There’s no reason to even consider SmartThings for this particular use case.
JMO
4 Likes
tgauchat
(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy)
#436
It would be interesting to figure out what the “least expensive” but entirely reliable system would be…
It may be but I don’t trust it. I also suspect things will be getting worse as more “christmas” hubs are added. I do have to say that my system is reliable when it come to “Actions” and not “events”. for example my Alexia control of lights using the ST hub is pretty reliable.
So I can buy a Aeon Panic Button for $41.95 and an alarm for $49.95. Hmmm. I could use the alarm for other things…
How reliable is the alarm, any opinions.
Seems $25 for the other system is a no brainer but the ST system could also perform other actions like sending text messages, turning on lights etc. It would be nice to have more control.
Unfortunately my wife has seen ST fail too often and says “No” to that solution. No WAF (Wife Acceptance Factor)
I wouldn’t trust ST with anything that matters. All too often they went in without notice or predefined expectations and wrecked things up in an effort to fix an issue. It will be few weeks until platform is stabilized and they get into the rhythm of being more mindful of their existing customers. They are showing progress but still too reckless to be trusted with someone’s life.
Not sure if it’s the right thread for such question but thread title tell me to do it here—
Yesterday, something weird happened unexpectedly. I have V2 hub that has 1 iphone and 1 Android registered as a presence sensor. SHM uses presence sensors to auto arm and disarm. Yesterday around 5:30 PM when we both were out, we got an intrusion alert with garage door open msg. I was able to close the garage door (NGD00Z, with your deice type Ron…Tx for that) remotely with ST app. Later on upon looking at the logs, it seemed an APP COMMAND was issued to open the garage from SmartTiles. However I wasn’t able to figure out which device would have sent that command. I suspected it may have somehow happened while my phone got unlocked in my pocket but that’s very unlikely though. This was a very scary event and fortunately all other doors were locked and I didn’t receive any motion event in the garage but I would like to get more details on how it happened. For now I am looking for a help to identify the device from which command was sent to open the garage door. I already looked into logs and device ids in log didn’t seem to give any info.
Here are the logs–
2016-03-20 4:55:44.996 PM PDT
18 hours ago DEVICE door closed Garage Door door is closed
2016-03-20 4:55:32.652 PM PDT
18 hours ago DEVICE door closing Garage Door door is closing
2016-03-20 4:55:25.701 PM PDT
18 hours ago APP_COMMAND close
2016-03-20 4:53:21.371 PM PDT
18 hours ago DEVICE door open Garage Door door is open
2016-03-20 4:53:13.931 PM PDT
18 hours ago DEVICE door opening Garage Door door is opening
2016-03-20 4:53:07.255 PM PDT
18 hours ago APP_COMMAND on
And
rawDescription (Smarttiles_Name) sent on command to Garage Door
smartAppId 0bcxxxxxxxxxxxxxxxxxxxx97e2518
smartAppVersionId 16exxxxxxxxxxxxxxxxxxxxxxxxxxfb2e55c
The system has been unstable for about a week, including database corruption and random mode changes. It seems likely that’s what you ran into unfortunately.
Thanks for your quick reply, I had been reading that thread but I just want to make sure of what/how it happened. Looking at the log it was obvious that an “on” command was sent from Smarttiles APP…is there any way to determine from which device ?
This kind of event is very scary and if it’s system error then that’s really is a big issue and I hope that’s not the case :(.
It is really a big issue, and the company is treating it as such. Database corruption means Device control messages can be lost, duplicated, or even just randomly sent.
As to your specific smarttiles question, maybe @tgauchat has the answer.
Yup… and a prime reason I don’t let SmartThings open/close my garage door, or adjust thermostats to keep pipes from freezing (or any critical task really). Monitor yes, control no.
The reason to have ST Control the Garage Door is…If the door was left open by any of us (happened twice) ST rules are set to close the door automatically after 2 min (when the door was left open and no motion detected in garage for 1 min)…so it does save us from human errors as well…
I get it… had mine doing the same thing initially. But couldn’t rely on it. Now it just texts me if it is left open, and I use the MyQ app to close it (which works 100% of the time).
Don’t get me wrong, I’d luv for all this to work… but I have been here two years and it’s always sumthin’.
Lol… I am completely with you…As much as I want to rely on this system…It disappoints me…:-(. Unfortunately I don’t know if NGD00Z have any app… so for now I am stuck with ST…Too bad
tgauchat
(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy)
#447
Unexplained (poltergeist !) Events and/or Commands are very concerning, so I’m happy to help.
All browsers/devices using the same SmartTiles child dashboard instance URL will show up as the same SmartApp Installed Instance ID in the log, so there is no way to tell which one.
If you believe there is any chance that the Access Token of that URL has been compromised, then go into the configuration pages for the SmartTiles SmartApp (using regular SmartThings mobile App), and select the “Preferences / Security / Tap to revoke Access Token” option. The child dashboard / SmartTiles instance will remain installed with the same ID, but the Access Token portion will be changed and can be fetched like the original from the View Dashboard - URL option.
As we may be drifting off-Topic; please email me with any questions: Support@SmartTiles.click
SmartTiles has no code that would intentionally or unintentionally issue rogue Commands, but we cannot rule out the possibility of “ghost events” that are inexplicably prevalent from time to time on the Platform (e.g,. if the database or messaging subsystem is overloaded, I would personally assume that there is a possibility that an old “on” command could have been re-queued and replayed, or played from a delayed request the seemed to fail earlier). I certainly hope this is not a common scenario, because delayed queued commands could original from any SmartApp or even the standard mobile App.
(SmartThings’s database is an “eventually consistent” model, which implies that database writes are permitted to be delayed indefinitely; I have no idea at all if this increases the likelihood of the delayed command / delayed event scenario described above, or any of the poltergeist activity in this Topic.)
Thanks Terry. I understood the possibility of delayed command, but in all likelihood we generally open the garage door with it’s own remote and reason for installing a zwave controller was only to get the door closed automatically in case it was left open accidentally…so I am certain that knowingly command was not issued. Now the two remaining possibilities are 1. My phone (Samsung Galaxy S5) juggled while in my pocket (I don’t have it set to lock, it’s swipe to unlock so it does happen pretty frequently that it gets unlocked unnoticed and start doing weird things) and issued the command via smarttiles web interface. 2. Something went wrong with ST system.
I was hoping that somehow it might be possible to find out from which device the command was originated. Looking at the log for Door opening event, I figured it’s the smarttiles app that was used to issue on(open) command…
rawDescription (Smarttiles_Name) sent on command to Garage Door
smartAppId 0bcxxxxxxxxxxxxxxxxxxxx97e2518
smartAppVersionId 16exxxxxxxxxxxxxxxxxxxxxxxxxxfb2e55c
Now if I can find out from which device, that could eliminate at least a few doubts…
Sorry, so you already mentioned that there is no way to find out from which device command was issued…that sucks…and seems horrible for security purposes.
tgauchat
(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy)
#450
You can choose to use a unique/distinct child dashboard on each of your devices, and then they will have a distinct “smartAppId” which will show up in the log and match the AppId that is embedded in the URL used on each device. SmartTiles v5 allows the creation of 5 unique/distinct child dashboards, each with its own URL and secret Access Token.
Version 6 of SmartTiles (under development, no ETA) will have a more granular security model and your idea of somehow tracking individual browser/device identification will be feasible and considered.
For Version 5, @625alex and I decided not to attempt to identify individual browsers as that would require a tracking cookie or other methods generally considered to lead to possible privacy violations. Our general policy is that SmartTiles should be tracking as little information as necessary, especially any information that could, without a lot of care, turn out to allow the identification of individuals somehow.
(Version 6 deals with a lot more data and inherently more tracking and diagnostics will be possible. We’ll have to have an easy to understand and enforce published Privacy Policy for this.)
Although The use case that initiated that topic is about someone who fell and blacked out, all of the commercial solutions mentioned for that problem also have a manual button that can be pressed to initiate the alert on demand, so they also solve your problem.
!) Events and/or Commands are very concerning, so I’m happy to help.