We had an incident over the weekend where it appeared a bandwidth restriction was put in place by our infrastructure providers.
After thorougher investigation the culprit was a problem with the machines NIC.
During the performance impact we diverted the bulk of our traffic to secondary servers, which ensured that billing and critical systems remained functional.
We are marking this as resolved as the issue was rectified on Saturday and we have observed stability since. However we will be paying close attention for a recurrence of this problem, if this does happen more severe action will be taken.