The problem is actually that it loops too fast: It sends as many loops as it can according to the connection speed. You could have one ezstream instance streaming to one fallback mount and specify that fallback mount for all other mounts. You don’t need one instance per mount.