Hi — the image service has been restarted, and seems to be working properly and rebuilding the cache nicely.
The issue was that one of the image processing servers was unstable. This has happened a few times recently, and whilst it is an easy fix, we currently don’t have a good and timely mechanism in place for warning us that it is occurring.
This is a far from ideal state of affairs, and so we are currently working on a reporting system that will initially warn us in good time, and eventually fix any problems as they occur.
In the meantime, we’re sorry for any inconvenience caused, and we’ll be monitoring the image service to make sure all’s well after the reboot.