Mmm no that’s not normal ^^
Everything should work… If you have more information, I’m interested 
It needs to be fixed.
Indeed, in Gladys 4, the directive is that devices should never be added automatically: it’s always the user’s choice, and this choice must be made in the UI.
You can see how other services like Philips Hue do it:
-
The service automatically detects devices, it stores them in a list of devices in RAM.
-
When the user arrives on the service page, they see that new devices have been detected, and they can choose to add them to Gladys, or not.
-
When the user clicks on « create » the device, it’s the UI that sends the request to the backend to create the device. Thus, the user gets visual feedback if something goes wrong. (a device already exists with this name for example)
In Gladys 4, we avoid everything that happens in the background and could go wrong, because if an error occurs, the user has no feedback. (And no, « check the logs » is not a solution, in v4 we consider that the user should not have to check the logs, like any product)