Aggregation Pipeline and Shard collection in MongoDB (Pipeline and sharded collections)

In MongoDB aggretion, if there is a matching shard key in the pipeline, then this pipeline only runs in matching Shard, in the previous (3.2), pipeline is diverted, and finally by primary shard merge;

In sharded collections, aggragation must run in multiple shards, and if this operation is not requested to run into primary shard, these operations will be routed to a random shard to merge the result, This eases the pressure on the primary Shard database to overload, $out and $lookup requests are run into the primary Shard database.

Aggregation Pipeline and Shard collection in MongoDB (Pipeline and sharded collections)