There is a minor opportunity for optimizing the mongo call(s) for archiving: we should include `{ inS3: false }` in the queries instead of filtering locally. It's been like that for years now, so it can stay a little longer.