A couple of our cache servers were being slow to serve requests (particularly on the /3 Zhou cluster). We've sorted that out and it should be performing normally again.

Resolved... or not!

Performance was still pretty slow. So I've gone in and done some database work and also set up a twice weekly 'optimizing' job that will clear out old pages on a regular basis.

Cache performance looks like its running fine right now. And I'll check it again a couple of times during the day.

This should only be affect /3 Zhou cluster customers.