The following section of VES spec is ambgious and should refined to provide specific requirements for how this is to be handled.

The sequence number for each Fault event is currently set to
0 on a raise and 1 on a clear. We could change that so that
each Fault event is given a new monotonically increasing
sequence number whether it is a raise or a clear if that is
helpful (which is reset to 0 if the VM restarts) but they