We identified a problematic disk which has been replaced. All systems are back to normal.
We have identified a performance problem with one of our databases affecting a subset of customers.
17:37 UTC - 20 February 2015