The global latency stats Server Status section currently aggregates all the latency information known to the server.

We have a need for the same information except aggregated by collection.

The actual command: db.serverStatus().opLatencies.

Often I have a $lookup and the next stage is a $addFields { $first: "" }} because I know the lookup will return only one entry. It'd be nice to have this option directly in the lookup, so it'd return the first object instead of an array. Thanks!

As we add new nodes, it requires us to update the URI. Would like a way to have this automatically update.

If the idea that data is stored within a region (via zone based sharding) and they are "just" replica sets, why can't we just restore a single shard? Right now we have to restore the entire cluster.

Allow Mongodb to access an external database for:
1.select identifiers on an external rdbms (ex: postgres, sqls, oracle, ...)
2.use these identifiers to perform aggregations in Mongodb

Currently, it is necessary to use a third-party tool that connects to several data sources to perform these actions.
The idea is to be able to perform these "joins" only in Mongodb CLI.

Regards,
Jerome

A build in ARM support (Apple M1) for Alpine Linux (or Debian) is needed to run the mongocryptd process. Currently the official node docker images are available only in Alpine and Debian.

To improve network security, please create an option to Deny specific Network Adresses to MongoDB Cluster.

Change streams currently return only the new version of the documents.

If there were an option that could be set on the collection for the change streams to also return the old version of the document, it would simplify any business logic that requires the old version.

SQL based option to create view using multiple collections or databases

CHoose an oplog event to set the resume token for, instead of needing to run the change stream for system to create a resume token

Why: Concerns about many stream re-reading a full oplog if failure before finding first event

Would be a great great feature if we can create a VPC between ATLAS/MongoDB/AWS and Heroku/AWS.
Thanks.
Adailton Silva

When cloning a Database it would be helpful if User/Roles could be cloned directly as well and carried over.

MongoDB has the great future of Workload Isolation based on node tags: we can add nodes tagged for analytics or reporting and then configure the same tag for this type of connections so that they do not interfere with the main operational workload.

But it is the responsibility of the BI/Data Scientist/Analyst to properly configure the connection: it cannot be enforced in a centralised way. It would be great if MongoDB permissions could be "tag aware", and allow access for a particular user only to some nodes of the cluster.

Suppose I have collection of documents { entityId, entity }, where entity is names of other collections I would like to make aggregation with lookup and be able to pass entity value to $lookup from and as

db.collection.aggregation([{ $lookup: { from: '$entity', let: { entityId: '$entityId' }, pipeline: [{ $match: { $expr: { $eq: ['$_id', '$$entityId'] } } }], as: '$entity' } }])

As of mongo v4.4.4 that seems not working.

As a DBA or Data Architect, I want to be able to enforce data governance policies like schema control. Users of MongoDB cannot change the schema however DBA or named individuals have the ability to change the schema.

geoIntersect queries on spherical geometries are not adapted to our use case. Our need is to retrieve a list of tiles intersecting a given area of interest (AOI) in a flat 2d geometry.

Classical geo-spatial databases (PostGIS and Oracle Spatial) support this kind of approach at no extra cost and we would expect the same from MongoDB.

In MongoDB , using a 2dsphere index forces us to implement a workaround adding an artificial margin to the AOI and making a post-processing to remove false positives. See attached slides for explanations.

Best Regards

I would like to perform a "contains" query on a single text field in a case insensitive and diacritic insensitive way. I would like the query to have the performance of the field being in a text index by itself (assuming that this gives the maximum speed).

It is usually the case that the text index already exists for multiple fields. This means that a text search then has to be combined with a complicated (to make it diacritic insensitive) regex expression to find the correct results.

https://developer.mongodb.com/community/forums/t/query-for-field-containing-a-string-case-in-sensitive-and-diacritic-in-sensitive/14814/2/

When porting an application backend from an RDBMS to MongoDB, we've spoken to two people who've are looking for a way to specify a limit on the number of documents in .updateMany(). I understand the behavior cannot be defined on a sharded cluster, but if we had a way to do this on an unsharded collection, that would help when dealing with these teams.

Currently, the $addToSet for upserting to an array checks for object equality (when the array elements are objects). We sometimes need to check for uniqueness based on a specific field of the object rather than the object as a whole, and the only way to do this is to force single-threaded operation and to use $pull and $push.

If there is a way to specify the field to check uniqueness for the $addToSet , we would be able to perform writes normally.