Video doorbell calls to every phone in a home

Your company makes video doorbells, and 1.5 million are installed. When a visitor presses the button, every phone in the household should ring like a phone call. A household has up to 6 phones with your app. The first person to tap Answer talks to the visitor on live video. Your media service already connects one phone to the doorbell's video once it is told which phone.

Phones should start ringing within 3 seconds of the press. The other phones should stop ringing within 2 seconds of the answer. If nobody answers within 30 seconds, the doorbell asks the visitor to leave a message. To ring a phone whose app is closed, you send a push through Apple's or Google's push service. It usually arrives in under a second.

Doorbells are pressed about 8 million times a day, and the evening peak is 3 times the average rate.

What happens between a press and the first answer, and how do the other phones know to stop ringing?