We've had to restart the databases multiple times today because of server freezes. During the freezes, users would have encountered error pages when trying to access their blogs.
We're planning on a number of improvements this week to address this very serious situation. First, we will be pushing new code to both gather information on these freezes as well as revise some features to put less strain on the database. Second, we will be effectively doubling the number of machines used to handle the db workload.
I will update status with additional information our progress toward these goals.