Mobile presence has been rolled up in the concept of ‘Member Location’ in the app. For some reason that I don’t think anybody has had a bash at guessing at, it seeming so odd, not only have the actual mobile presence devices been hidden from view, but you can’t see the current Presence and Occupancy states anywhere at all. The same applies to the Linked Places that have been added to mobile presence on some devices.
However other devices using the Presence capability, such as the Arrival/Presence Sensor fobs, do still appear as devices.
Using phones as presence sensors worked pretty seamlessly for me prior to migration. However since migrating, the lag time has made it unusable.
I will arrive home, open the door and set off the alarm because the presence status has not yet updated to indicate that I have arrived home.
I’m also having issues where the “I’m Home!” routine runs before we leave in the morning, as if our phones just arrived when in reality we’ve been home since last night. Both Apple and Android phones are set to allow SmartThings to access Location all the time.
Looking for a way to get everything working the way it was again.
I’m using cell phones as presence sensors. The Classic ST app had phones as things that displayed when they were present or not. Since the new app does not display the actual cell phones, I was advised to create simulated presence sensors in the IDE which I did. I have a webCoRE piston which sets the simulated phone sensors state to arrived or departed based on the real cell phones. In my experience (as others have also mentioned) the reliability seems to be questionable. But it’s the only thing I have at the moment.
However, now to my question. I was looking at the app history for one of the simulated phone sensors today and it correctly shows the phone left home at 9:16 am and arrived back at 1:00 pm. The event log in the IDE for that sensor matches to the minute. The log for the webCoRE piston matches to the second with the IDE. BUT, the IDE event log for the actual phone is missing the departure time altogether, and it shows the phone arrived back home 2 minutes and 32 seconds AFTER the webCoRE trigger fired. In fact, the actual phone’s log shows a departure time from yesterday, no arrival time back yesterday, and no departure time from this morning. So it is missing 2 time stamps, and a third one is late. How is this possible? Is this just a bug in the ST IDE logging? Something caused the webCoRE trigger to fire correctly. I just don’t get it.
I am having a similar problem. I also use my iphone as a presence sensor but through the life360 app. In the past couple of days I’ve noticed that my Webcore automations that rely on my life360 presence are not working.
What does the history show about the occupancy attribute? That seems to relate to Wi-Fi connectivity but ST have never commented on exactly what it does.
Because it makes no sense to show occupied when you are not present, the device integration will always set present immediately before occupied if it isn’t already. I don’t normally see that happen as my phone normally detects me nearby a couple of minutes before I am in Wi-Fi range.
I do often see the reverse happen. On departure I see unoccupied set immediately before not present`.
I use an Android phone and all that behaviour is visible in the logs. I am just wondering if, for whatever reason, you only see half the story in your logs.
Until you mentioned it, I never paid attention to the fact that, in the IDE and in the ST app history, the actual device’s name and value columns are name=occupancy and value=(occupied vs. unoccupied), while the simulated presence sensor is name=presence and value=(present vs. not present).
Here’s an scaled down version of my IDE log:
S8 = Samsung Galaxy S8
S8Sim = Simulated presence sensor for the S8
BB = Blackberry Keyone (yes some people refuse to give up their Blackberry!)
BBSim = Simulated presence sensor for the BB
Device Date Name Value
S8 2021-02-07 4:12:12.000 PM EST occupancy unoccupied
S8Sim 2021-02-07 4:12:12.000 PM EST presence not present
BB 2021-02-07 4:14:19.000 PM EST occupancy unoccupied
BBSim 2021-02-07 4:14:20.000 PM EST presence not present
S8Sim 2021-02-07 5:17:24.000 PM EST presence present
S8Sim 2021-02-07 5:17:34.000 PM EST presence not present
BBSim 2021-02-07 5:18:46.000 PM EST presence present
BB 2021-02-07 5:21:59.000 PM EST occupancy occupied
S8Sim 2021-02-07 5:34:07.000 PM EST presence present
S8 2021-02-07 5:41:15.000 PM EST occupancy occupied
So the departures seem to be correct. It’s my arrivals that are messed up, which also screw up the pistons I have that rely on phone presence.
For example, my S8. The simulated presence shows me arriving at home, then leaving 10 seconds later. There is no entry for the physical phone to even trigger the simulated one’s arrival (nor departure). I wonder if my departure is caused by the phone connecting to my home wifi? Then the simulated phone shows I arrive again at 5:34, when the physical phone does not arrive until 7 minutes later at 5:41. The simulated phone should be based on the physical one.
And as for my wife’s Blackberry, at least hers doesn’t show arriving/departing/arriving. But it does have the simulated one arriving 3 minutes before the actual phone.
Since I have webCoRE pistons that turn things on or off based on departure/arrival, these fail or more likely give me erroneous results because, in my case. the system still thinks I’m away. This is a less than ideal situation.
Here is what my A20e shows from a brief walk this morning (IDE Events, latest first).
[2021-02-08 11:25:20.000 AM GMT] occupancy occupied
[2021-02-08 11:23:00.000 AM GMT] presence present
[2021-02-08 11:03:31.000 AM GMT] presence not present
[2021-02-08 11:03:31.000 AM GMT] occupancy unoccupied
As you can see, on departure the occupancy was set at the same time as presence, so that means the phone reported not present and the “handler” faked the unoccupied to match it. On the way in present was detected long before occupied so both are real.
I do mirror the mobile presence onto a virtual device, purely so I can see stuff in the mobile app. So just for completeness:
[2021-02-08 11:25:20.000 AM GMT] occupancy occupied
[2021-02-08 11:23:00.000 AM GMT] presence present
[2021-02-08 11:03:32.000 AM GMT] occupancy unoccupied
[2021-02-08 11:03:32.000 AM GMT] occupancy unoccupied
[2021-02-08 11:03:32.000 AM GMT] presence not present
The mirroring uses a Rules API rule. The duplicate occupancy on departure is probably caused by the presence and occupancy events firing the rule at the same time. I thought I could make sense of it this morning. Now, not so much.
Are your phones really not showing presence in the event logs and history? Regardless of that the departures look similar to mine with not present being detected and unoccupied being set to match it. The arrival on the BB (still my favourite ever phones) looks normal - present being detected and occupied turning up a little later - but the S8 does seem to have thrown a wobbler.
In other event traces of ‘bouncing’ arrivals you often see the first present coincide with occupied as if the phone connects to the Wi-Fi (or whatever the trigger is) first and the presence is ‘faked’ to match. However the phone then immediately decides it is definitely not present and the unoccupied is set to match it. Later it decides that yet it has indeed arrived and the phone is present again. I’m not sure what happens to the occupancy in that scenario. You don’t seem to have that early occupied though, you just seem to have a very bewildered phone.
I don’t know what causes the bouncing. As I’ve often seen Google’s location history doing things like jumping around three places in London in an instant, or indeed between London and Helsinki, I’ve often suspected there is a lot of dubious Wi-Fi location information about, especially mobile Wi-Fi that hasn’t been identified as such. That can’t help.
I believe that is what I am doing with the simulated presence sensors I had to create in the IDE after migrating from the Classic app to the current one because the current app doesn’t display real phones as “things”.
Where is this? I am doing my mirroring using webCoRE. It, or at least I thought it was supposed to do, was get a triggered event (occupied/unoccupied) from my phone, then make the simulated one present/not present. That way I can see the status of the phones.
I assume you mean in the ST app itself? If so, yes it does show up in there. However, 1) the IDE shows greater detail, like down to the seconds vs. minutes in the app, which is helpful to see which event came first like I am experiencing with seconds of delay, and 2) I’m also finding the app isn’t always sync’ing completely. Just today the app history (the full log) has no events between 1:54 pm and 5:39 pm. Yet the apps log for my garage sensor shows multiple temperature changes in that time range, as well as the garage door opening and closing around 2:20 pm. Why these don’t show in the big history log is strange.
And yet some days it works just fine. Very aggravating.
HOpefully what I have to say may help someone. In my ST account my Iphone has never had any problems. With, without webcore, whatever, never had a problem ever. My wife’s however is a dfferent story. About a year ago after switching to the new app, her Iphone 6 quit all location reliability. Whether in the app, out of the app, ad nauseum… ( usually opening ST app on her phone would get a fresh location update adn trigger the app )
At the time I fixed the issue with webcore and some pistons. It worked perfectly. Fast forward to present. A recently purchased Iphone 12 and shutdown of Iphone 6 did not work reliably either. Previously a ST ticket ( 2021 ) answer they told me that originally it was because her phone was too old and I should upgrade her phone. that I fixed at that time using webcore. But then once upgraded, Mar 2022, the new phone did not work correctly either…
Careful thought about the on going problems led me to consider HER SAMSUNG ACCOUNT… since the phone device is ultimately tied back to each individuals account… About a week ago I accessed her Samsung account ( Samsung US | Mobile | TV | Home Electronics | Home Appliances | Samsung US) and noted that it showed 2 iphones as her devices. Even though one of those had been turned off… SO I deleted ALL phones from her SAMSUNG ( Not smartthings) account and then deleted her from ST and proceeded to re-invite and re-enter her into ST. Since removal of the extra phone in her Samsung account, there have been ZERO issues. I’m no longer using webcore, since that is going away soon anyway. I am using a mobile presence sense to make a virtual switch as a device handler in ST, but I bet I do not even need that. In any case, If you are having immeasurable fails with your geo fence reliability, check on your SAMSUNG account and see how many phones are in there…Maybe it works for you too