Some customers like ComboCurve, ServiceTitan, & LucidMotors have requested Full Cluster change stream support. In many of these cases, many databases must be captured, and it is non-tenable from both a cost and operational management perspective to have to run a processor for each database to be collected. 

The proto could look like this, which would capture ALL of the change stream for the connection. Just simply not adding database or collection to the $source. 

{$source : {
           connectionName : 'jsncluster0',
           config : { pipeline : '...' } 
           }
}