Similar to Spring Data, can connection pooling be done on Firestore/CollectionReference. The quickstart shows steps to create connections for reading/writing, but could not locate if these Java objects support concurrency.
Not: referring Database Connection pool , not http pooling.
Connection pooling allows you to re-use objects that are expensive to create. The Firestore SDK already re-uses such objects behind the scenes, and there's neither a need nor a way to influence this.
The Firestore object itself is already cached by the SDK, or at least its expensive internals are. A CollectionReference object is lightweight wrapper for a path in the database, so it's cheap to create.