No, it’s occurring in the middle of the list (though the changed url returns no results, so becomes the end of the list for that iteration)
The first occurrence happened at offset 2500, and the second occurrence happened at offset 600 (so offset 0-2400 worked as expected, and 0-500 worked as expected)
To note, it also fixed itself between the two error scans, where it got a full good scan done (~15000 offset max, but this changes on the time of day, usually around 11000-24000)