I started a thread over in the Coffeehouse asking if anyone else was seeing this, but apparently not.
I keep getting request timeout errors for C9 for periods of up to 45 minutes:
29 Dec 2009: 08:15 to 09:00 GMT
29 Dec 2009 15:45 to 16:05 GMT
30 Dec 2009 08:27 to 08:44 GMT
I shall keep recording these timeouts for a while to see if there's a pattern -- is there a maintenance/backup schedule twice a day that is slowing the server(s) to a standstill?