I've seen several cases where match_all queries were being used inside percolator queries,
because these queries were created generated by other systems.
Extracting these queries will allow the percolator at query time in a filter context
to skip over these queries without parsing or validating that these queries actually
match with the document being percolated.