Amazon Neptune · Rate Limits

Amazon Neptune Rate Limits

Amazon Neptune has two distinct API surfaces. The control-plane (CreateDBCluster, etc.) is throttled per-account/region under the standard AWS API throttling envelope. The data-plane (Gremlin/SPARQL/openCypher endpoints) accepts queries up to the configured query timeout and concurrent-request limit on each cluster, scaling with instance class. Use exponential backoff with jitter on ThrottlingException and ConcurrentModificationException.

4 Limits Throttle: 400 Quota: 400
Rate LimitingDatabaseGraph DatabaseNeptune

Limits

Neptune control-plane API throttling account/region
varies
see Service Quotas console for Neptune
Standard AWS API throttling. Limits depend on action.
Concurrent open queries (data-plane) cluster
concurrent_requests
varies by instance class
Approximately 8 to 80 concurrent queries depending on instance type; configurable via DB cluster parameter group.
Default query timeout cluster
seconds ยท second
120
Configurable via neptune_query_timeout parameter (max 2,147,483,647 ms).
Clusters per account/region account/region
count
40
Soft default; raisable via Service Quotas.

Policies

Backoff with jitter
AWS SDKs default to standard retry mode (truncated exponential backoff with jitter, max 20s, 3 attempts).
Read scaling
Add up to 15 read replicas per cluster to absorb concurrent queries.
Long-running queries
Use the Neptune Status API (/status) and explicit /gremlin/status endpoints to monitor and cancel long-running queries.
Quota increases
Most quotas are soft and can be raised via Service Quotas.

Sources