On 4/15 around 10pm we saw packet loss for our URLFetch service which
may have caused URLFetch requests to time-out and return errors. The
cause of the network problem was tracked down and rectified about an
hour an 45 minutes later, restoring normal service.