Resolved - From 0030 until 0745, the website front-end was unavailable, and users will have seen requests timing out. This was caused by an uncommon failure of two servers within 30 minutes of each other, both due to one of their disks filling up with log files.

We are investigating why our monitoring systems failed to alert, as they are set to do, several days before a disk is likely to run out of space.
Mar 4, 00:30 GMT