Memgraph · Rate Limits

Memgraph Rate Limits

Memgraph is a self-hosted database accessed over the Bolt wire protocol rather than a metered public REST API, so it does not impose vendor-side RPM/TPM request quotas. Throughput and concurrency are bounded by the operator's own configuration and hardware - the Bolt session/connection limits, the query execution thread pool, the available memory for the in-memory graph, and any storage/transaction settings. There are no published per-account rate limits; effective limits are determined by deployment sizing. Specific configuration values are not reconciled in this artifact.

Memgraph Rate Limits is the machine-readable rate-limit profile for Memgraph on the APIs.io network, conforming to the API Commons Rate Limits specification.

It captures 4 rate-limit definitions, measuring connections, queries, memory, and requests.

The profile also includes 2 backoff/retry policies defined and response codes documented for throttled.

Tagged areas include Graph Database, In-Memory, Cypher, Bolt, and Real-Time.

4 Limits Throttle: not applicable (Bolt protocol; no HTTP 429 quota on the query interface)
Graph DatabaseIn-MemoryCypherBoltReal-TimeRate LimitingQuotasThrottling

Limits

Bolt Connections / Sessions deployment
connections
operator-configured
Concurrent Bolt sessions on port 7687 are bounded by server configuration and host resources, not a provider quota.
Query Concurrency deployment
queries
operator-configured
Bounded by the query execution thread pool and available CPU/memory; tune via configuration flags.
Memory Capacity deployment
memory
operator-configured
As an in-memory database, the working graph must fit within configured/available RAM; query memory limits are configurable.
Metrics HTTP Endpoint deployment
requests
operator-configured
Enterprise metrics endpoint (default 0.0.0.0:9091) is intended for scrape by Prometheus; no published rate cap.

Policies

Operator-Controlled Limits
All effective limits are determined by the self-hosted deployment's configuration and hardware, not by Memgraph as a vendor.
Capacity Planning
Size memory and CPU to the graph and concurrent workload; use HA/replication for scale-out read throughput.

Sources