sorry, missed the other post.
The outage the other day had to do with a router failure (i'm not sure which side as it was not specified in the error report). I do not know why things didn't fail over to another route. My guess is that when the router failed things snow balled and affected other programs/links. I understand work is underway to improve the reduncies and failover mechanisms.
The exchange released new software for Globex which I know created a few hiccups over the past month or two. They seem to be ironed out now. Ofcourse, by me stating this murhpy's law will rear its ugly head.
All I can say about satisfying the desire to know why a connection goes down or why certain decisions have been made, is that there is an internal discussion going on. I believe it will lead to more information being passed to clients on a timely basis.