Prisma · Rate Limits

Prisma Rate Limits

Prisma Data Platform (Postgres + Accelerate) does not publish a per-second rate-limit table. The documented quantitative bounds are response-size and connection-related — Accelerate flags retrievals over 5 MB as significantly slowing the application (error P6009), and the default Postgres connection limit is 10 unless `connection_limit` is configured. Plan-level monthly operation quotas (free / starter / pro / business) effectively serve as the consumer-facing rate ceilings.

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

It captures 3 rate-limit definitions, measuring operations, varies, and concurrent_requests.

The profile also includes 3 backoff/retry policies defined and response codes documented for throttled.

Tagged areas include Database, ORM, Postgres, and Rate Limiting.

3 Limits Throttle: 429
DatabaseORMPostgresRate Limiting

Limits

Plan monthly operations (Postgres + Accelerate) account
operations · month
plan-defined (Free 100k Postgres / 60k Accelerate; Starter 1M Postgres; Pro 10M Postgres; Business 50M Postgres)
Once a plan's included operations are consumed, additional operations bill at the plan's per-1,000 overage rate rather than being throttled.
Accelerate response size query
varies
5 MB recommended ceiling per response (P6009 P6009 — slows app significantly above this)
Accelerate documentation flags retrievals over 5 MB as the practical response-size bound; query timeout is also configured but the exact value is not published.
Postgres connection limit database
concurrent_requests
10 (default)
Set by the `connection_limit` parameter on the database URL; raise based on workload and plan capacity.

Policies

Plan-Quota Throttling
Hard request throttling is not the primary lever — plan-level monthly operation quotas cap usage, with overage billed at the plan's per-1,000 rate.
Connection Pooling via Accelerate
Use Prisma Accelerate's connection-pooling and edge cache to reduce direct Postgres connections; cache tags (Pro/Business) let consumers invalidate specific cached queries.
Response-Size Discipline
Keep individual query responses under ~5 MB; paginate large result sets to avoid the P6009 slowdown.

Sources