We have a hierarchical structure, stored in a Mongo collection, in which essentially any record can be a parent of another record up to a depth of four. Mongo advocate embedded documents as a good design practice yet performing $graphLookup against records with a sizeable child set often causes the 16MB document limit to be exceeded. This seems to go against the entire ethos that related data should be embedded as we often fall foul of this restriction. Surely doing some sort of pre-projection would be beneficial so that only the fields of interest are included into the final aggregated document. There appears to be a ticket in place related to this https://jira.mongodb.org/browse/SERVER-38560) related to this but it is currently unresolved. Also, we are currently on Mongo 5 and would be interested if this feature (fix?) would make it into this version as a minor release or whether we would need to do a major version upgrade which presents its own problems. This is also fairly problematic for us, so some sort of timeline on when this could be expected would be useful, or if this is even likely to make it in at all? Thanks in advance for your help.