Could have sworn I already replied to this :f

The reason I have a hard time believing 1/2 are likely is because of the gap ranges. 7-8 times the sequence gaps have been over 1 million. There shouldn't be any kind of inserts on that level (6000 yesterday).

That said I looped 10,000 with the same kind of table structure (cluster index on some random column, non-clustered index on the identity PK) but never had a gap. Tried updates in the hopes it might have done something but nope. I guess I'll try a loop tomorrow starting with numbers around the ranges where the skips happened; just to rule out some weird dislike of the numbers by sql server).

It worked fine from 2006-2012 then suddenly started having issues in Jan 2013 so I'm hoping some application/code update was carried out around that time.

As far as I'm aware the only difference between a sequence & identity column is the scope (identity = single column sequence)?