Sorry for the delay all. I’ve been hesitant to say anything until we had confirmation, and I wish I could say “this will be resolved by -insert date here-” but it’s no secret we’re having a particularly hard time solving this.
With that in mind, the plan is to patch next week with additional telemetry capture so our developers can identify where and how exactly these connections are breaking down. What we have in place at current isn’t quite enough to diagnose an issue this complex.
I (and Lev, and Hedge) will do all we can to support our developers in solving this and provide regular updates to the community explaining where we’re at on this issue. I know blanket PR apologies are of no use here, actions speak louder than words and whatnot, but we really do apologise that this has been an issue for so long. 