Goal

Let a Spring Boot 4.x application use MongoDB-backed JPA (via the MongoDB Extension for Hibernate ORM)
with minimal, idiomatic configuration — a plain @SpringBootApplication, no knowledge of Hibernate,
JPA, or Spring Boot internals, and no manual bean wiring.

Requirements

• Activate and connect through standard Spring configuration (Spring's MongoDB connection setup plus the
  standard spring.jpa.* properties), not bespoke properties or custom annotations.
• Reuse the application's single MongoClient / connection pool — never open a second pool to the same
  cluster when the application also uses Spring's MongoDB support, and never take over the client's
  lifecycle.
• Honor the full spring.jpa.* surface (ddl-auto, show-sql, naming strategies, Hibernate/EMF
  customizer beans), as in any Spring Boot JPA application.
• Provide Spring Data JPA repository support without requiring a SQL DataSource.
• Stay completely inert when MongoDB-JPA is not intended — e.g. a SQL-JPA application that merely has the
  module on its classpath, or a Spring Data MongoDB application that has a MongoClient but no Hibernate
  JPA — with no interference to a SQL persistence unit.
• Default field naming so documents are not silently reshaped relative to plain Hibernate (no surprise
  camelCase → snake_case on a schemaless store); allow an explicit override.
• Fail fast with an actionable message when activated but misconfigured (e.g. no MongoClient available, or
  no database name resolvable).
• Offer an escape hatch for advanced cases (a dedicated client, multiple persistence units) without
  fighting the auto-configuration.

Compatibility

Spring Boot 4.x; Hibernate ORM 7.3 or later. (Spring Boot's dependency management may force an older
Hibernate from its BOM; until a Boot release ships 7.3+, the application overrides the managed version.)