Amazon Redshift · Rate Limits
Amazon Redshift Rate Limits
Amazon Redshift exposes three API surfaces - the Redshift control-plane (CreateCluster, DescribeClusters), the Redshift Data API (ExecuteStatement, GetStatementResult), and the Serverless control-plane. All follow standard AWS API throttling per account/region. Cluster-level concurrent connection limits and query-queue concurrency are enforced at the database level. Use exponential backoff with jitter on ThrottlingException.
5 Limits
Throttle: 400
Quota: 400
Rate LimitingData WarehouseRedshift
Limits
Redshift control-plane API account/region
see Service Quotas console for Redshift
Standard AWS API throttling envelope on management calls.
Redshift Data API account/region
see Redshift Data API quotas
ExecuteStatement, BatchExecuteStatement, ListStatements, etc. are throttled per account/region.
Concurrent database connections cluster
500
Default per provisioned cluster; configurable, raisable via support.
Concurrent queries (default WLM) cluster
50
Cluster-wide concurrency cap across WLM queues; tune via WLM configuration.
Query queue slots per WLM queue queue
50
Maximum total slots across queues is 50.
Policies
Backoff with jitter
AWS SDKs default to standard retry mode (truncated exponential backoff with jitter, max 20s, 3 attempts).
Workload management
Configure WLM (manual or auto) and short-query acceleration to spread query concurrency; use Concurrency Scaling to absorb bursts on Provisioned.
Serverless auto-scaling
Redshift Serverless absorbs concurrency by adjusting RPU capacity; bills are still subject to RPU-hour rate.
Quota increases
Cluster, snapshot, and IAM-role limits are soft and raisable via Service Quotas.