Fixed - foreground services are killed by the OS but there is a catch

[cj-marius-duna]

Android 12 and later versions are designed to terminate foreground services automatically after days or weeks.

09-17 08:02:23.129  2670  2818 W ActivityManager: Exception when unbinding service com....otm.v3/com....otm.vpn.service.FilterService
09-17 08:02:23.129  2670  2818 W ActivityManager: android.os.DeadObjectException

If you check the logs and your service was a STICKY one, the ActivityManager is automatically restarting it with the following kill reason:
ActivityManager: Rescheduling restart of crashed service com.cujo.otm.v3/.service.CujoService in 9967ms for mem-pressure-event - mem pressure is the reason in my case. Just follow the LogCat and find yours

This means that the onStartCommand has already been called, but with a null intent.
In this situation, you will need to manually reinitialize everything.

[Duna]

Spend 2 weeks for finding this behaviour. Thank me later

[petrnalevka]

Hello @Duna @cj-marius-duna .. in my opinion this is nothing special in Android 12.. this is how FGS (or in earlier Android simply Services) worked on from the well "beginning". Even when the the service was sticky system may restart the service any time when memory is getting low and you need to cope with that in the code... for instacne expect intent to be null or expect a flow when onCreate is not called in the lifecycle but only onStartCommand...