There's a great inside-the-workshop post on Ello's lessons in scaling
If you haven't read "5 Early Lessons from Rapid, High Availability Scaling with Rails" and you've had eng/ops experience, or are just plain curious as to how Web infrastructure works, it's a great piece by Mike Pack of Ello.
I see numerous familiar lessons myself, as well as hints as to some of the behavior, quirks, and events noted with Ello itself. Including, I think, the reason why @ello no longer has a follower count.
I've got a prediction which is that a number of the lessons learned here will themselves be the basis of further lessons learned...