Amazon Neptune
Fully managed graph database for data whose value lives in the connections between things.
Verified against AWS documentation · 2026-09-09
Key facts
- 01Stores nodes, edges, and properties — 'friend of a friend of a friend' is one traversal, not three JOINs
- 02Query languages: Gremlin and openCypher (property graph), SPARQL (RDF) — no SQL
- 03Lives inside your VPC, no public endpoint; storage spans 3 AZs, self-heals, auto-grows 10 GiB → 128 TiB
- 04Up to 15 read replicas share one storage volume; failover to a replica typically completes in about 30 seconds
Exam triggers
Phrases and scenarios that point here.
- "Highly connected data", "relationships between entities", the word "graph" → Neptune, every time
- Social network, recommendation engine, fraud ring, knowledge graph → Neptune
- Query has to hop several levels deep (who knows who, who owns what) → Neptune, not recursive SQL
- Same person hiding behind 5 accounts, 2 devices, 3 emails → Neptune (identity resolution)
Not this service
- JSON documents, MongoDB-compatible → DocumentDB
- Key-value or wide-column lookups at massive scale → DynamoDB
- Tables, joins, SQL → RDS or Aurora
- Full-text search and log analytics → OpenSearch Service
Use cases
- 01Fraud rings: five 'unrelated' accounts that quietly share one device fingerprint and two phone numbers — obvious as a graph, invisible as rows
- 02Recommendations: "people who bought this also bought…" is a two-hop walk from product → buyers → their other purchases
- 03Identity resolution / customer 360: stitching one human together from email, phone, loyalty card, and device IDs
- 04Blast-radius mapping: if this switch dies, which apps go down? Walk the dependency graph three hops out instead of reading a wiki page
- 05Knowledge graphs behind GenAI apps, so the model answers from real relationships ("which suppliers feed this factory") rather than guessing
How it's charged
- 01You pay for the cluster running, not per query — instance-hours plus storage per GB-month plus backup storage. An idle Neptune cluster is not a cheap Neptune cluster
- 02Two configurations: Standard adds a per-million-I/O charge; I/O-Optimized drops I/O charges but raises instance and storage rates — AWS suggests switching once I/O passes roughly 25% of your Neptune bill
- 03Neptune Serverless bills per second in Neptune Capacity Units (1 NCU ≈ 2 GB memory), scaling down when traffic drops — the fit for spiky or dev workloads
- 04Neptune Analytics is separate and in-memory, billed per m-NCU-hour (1 m-NCU ≈ 1 GB memory); a paused graph keeps its data at roughly 10% of the compute price
- 05Backup storage is free up to 100% of your cluster storage; Database Savings Plans (1-year commit) can cut Neptune costs by up to 30%
