I wanted to give an update on our progress in dealing with the database performance problems. We have recently deployed a new set of servers to help shoulder the increased load - these servers are significantly better hardware (more memory). As a result, they'll help increase responsiveness throughout the app.
We've also identified a specific set of queries responsible for the database freezes we've seen. By re-writing these queries, we've been able to get to the heart of the most serious problems we've been facing. We now believe that the databases are signficantly less likely to freeze.
We have additional work to do. For example, users with larger blogs are still seeing unacceptably slow load times when accessing their blogs from the dashboard. There are other improvements we will be making as well as a completely replacement of our current infrastructure in favor of a more robust datastore. However, we're optimistic that the situation has improved.
