- If you need a self-hosted graph database
- Neptune is an AWS-managed service; air-gapped or non-AWS hosting requirements point to self-hosted Neo4j or open-source alternatives.
- If you require strict Cypher feature parity with Neo4j
- AWS documents openCypher compliance limitations; if your application depends on specific Neo4j Cypher behaviors, validate compatibility or consider Neo4j.
- If cost predictability is the top constraint
- Neptune's pricing dimensions (compute, storage, I/O, backups, transfer) can make forecasting more complex than bundled pricing models.
- If you need multi-region active-active writes
- Neptune Global Database supports a single primary write region; secondaries are read-only until promoted.
- If you want an integrated graph-platform experience
- If integrated graph tooling, data science features, and plugin ecosystems are core requirements, consider platforms that emphasize that vertical integration.
- Standards/query-language limitations
- Neptune supports openCypher but documents compliance differences relative to other Cypher implementations; validate feature compatibility for Cypher-dependent applications.
- Protocol/endpoint constraints
- Some Neptune endpoints and API surfaces have protocol constraints documented in the product documentation (for example, specific HTTP protocol support for certain endpoints).
- Global Database constraints
- Single writer region; secondaries are read-only until promotion for write failover.
- Documentation gaps (community-reported)
- Some users report difficulty understanding how to use newer Neptune Data APIs based on the documentation available at the time of their evaluations.
- Roadmap items
- This research run did not compile a Neptune roadmap; AWS feature delivery is typically announced via service documentation and "What's New" posts.