I should have mentioned that there are errors in that log. On the day the job started to fail there are errors (16060) at around 5am that the vms on the server were paused because the volume on which the vhds are housed ran out of space. Space was added to the volume and the vms restarted. At around 6pm - which is when the job starts - there are errors (10137) that "no snapshots to revert were found for virtual machine '<VM name here>' (Virtual machine ID <GUID here>). There are nine of these errors (corresponding for the 9 vms) for a few days at the time the job starts (6pm) and then the errors stop.
After that there are only two errors and they are from the same day the above errors stopped and about two hours later:
(ID 19010)
The WMI provider 'Msvm_ServicingSettings.Version' is not registered in the CIM repository.
(ID 14096)
Virtual Machine Management service failed to start.
NOTE: The above service started soon after and is running ok
Thanks