Mongosync and dbCheck would benefit from being able to do a natural scan on a specific recordId range like so:

db.runCommand( { find: "coll", sort: { $natural : 1 }, filter: { $and: [ { $recordID: { $gte: 100 } }, { $recordID: { $lte: 200 } } ] } } ) 

For mongosync in particular, it would be optimal if this is done efficiently when reading from sharded cluster, e.g. if a recordId range does not pertain to a specific shard, that shard should not be a bottleneck on the query.