We’re seeing POLL HANGUP DETECTED more often in recent weeks (on FS 1.10.11 triggered in mod_verto.c) which is causing calls to fail even when they seem to be going fine. From looking at the source code, this event seems to fire when FS detects that the remote end of a verto web socket has closed. Yet this triggers a command on that socket to logout and our browser clients are receiving that and hanging up on users.
Is this event caused by something else that we can enable more logging about?
Or, if it is trying to detect a socket break, can we configure it to be more tolerant, perhaps by waiting longer to see if the connection is live?