@Chancleta 

Yes, I've seen it too, and am similarly sized.   FWIW, these switches are 802.11N and so unable to do MU-MIMO.   As a matter of technology then, they can't concurrently communicate effectively  CSMA/CA should minimize this conflict but outlying/weaker signal devices may still cause problems.  If so, they'll do it for all devices.

I've broken my routines down into three sub-groups of 12-15 each.  I call the routines vis Alexa and implement a 2 second (1 would have been fine) delay between each routine firing.  Having done that I no longer experience random lights being left on.

Of course using a good wireless router and/or access point helps as well.  If you're using something your ISP gave you, well...

Having the right number  of quality AX routers on unique channels using the same SSID, properly configured to "roam" devices above say 70db, would be a best practice.

Conversely, making many concurrent calls over to 802.11N devices via a single entry level router/AP would not be "best practice".

I don't know the TP-link router family well, but I consider an RT-AX68U the entry level for smart homes.   An AC version, if already owned one, would effective too. 

Luck to you,

-d

  @dafish 

thanks. Hoping tech support has some insights as well. 

Does anyone from tech support or dev team have any thoughts/comments? Please.

@Riley_S 

? 

  @Riley_S 

For the first question, I believe it happens mainly on dimmers but also on non-dimming switches. Below are the details of one dimmer which has shown the issue many times. However, it is sporadic and happens randomly across devices.

Device Info

Model: HS220

Hardware Version: 2.0

Firmware Version: 1.0.8

You also mentioned the Beta version of the app for the scenes disappearing issue. That probably magically went away so I will wait for the released version to deployed before updating.

@Riley_S 

The issue continues to occur.

Since my post, I have paid better attention to what devices it seems to happen with. It *may* be that I need to walk back the statement that it happens with switches. Every occurrence since the original post has been with dimmers only.

Thinking about this further, I believe this is also a SAFETY Concern and potentially very dangerous.

Imagine a scenario where a dimmer controls a light bulb and the bulb is burned out or removed. The only visual indication would be the LED at the switch. In this scenario the LED indicator shows that the dimmer is off (open, meaning no current to the load). In reality it would be live. This is very dangerous. Another possible scenario, with other implications could be with a device tied to an outlet.

Obviously in the above scenario, the statement can be made that people should trip the breaker and/or test with a multimeter, etc. That said, how many people ever do this when changing a bulb. Basically nobody. 

So beyond being a nuisance and "bug" in the dimmer functionality, there is the real possibility of someone using the LED indicator on the switch as the "test of current at the load." This could hurt someone.

Is there a way I can pull logs, etc. for the development team the next time this happens? I am very happy to be support the debug of this situation.

@Riley_S 

Anything from the dev team?

@Riley_S  

Hello. Over Christmas I observed the issue a few times again, all with dimmers. I even saw an instance of the opposite behavior... the switch showed "on" but the light was off. After toggling the light came back on properly.

Uninitialized varibale or some type of race condition in FW? Or perhaps there is something mechanically wrong with the design of these dimmers that has this issue pop-up periodically.

I am the only person reporting this behavior?

I am up to 4 different dimmers which have shown this behavior. Maybe more but did not keep track for a while.