The signaling server is where a lot of the real complexity comes from [1]. I even had to write my own from scratch for our production implementation.
[1] https://www.sococo.com/blog/webrtc-signaling-here-be-dragons
[1] https://www.sococo.com/blog/webrtc-signaling-here-be-dragons
https://www.html5rocks.com/en/tutorials/webrtc/infrastructur... before implementing that one?
Their ready-made signaling server suggestions are complete garbage (Signalmaster is buggy and incorrect).
The one thing that is not handled in any material I've ever read on WebRTC: how to sequence more-than-two users joining a session at the same time, and how to recover from errors and stalled connection attempts (where the latter is a significant impediment to the former).