specification: API Commons Rate Limits
specificationVersion: '0.1'
schema: https://raw.githubusercontent.com/api-evangelist/interface-research/main/schema/api-commons.yml#/$defs/RateLimits
provider: Azure Synapse Analytics
providerId: microsoft-azure-synapse-analytics
created: '2026-05-04'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-04, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-05'
reconciled: true
tags:
  - Analytics
  - Apache Spark
  - Big Data
  - Data Integration
  - Data Warehouse
  - ETL
  - SQL
  - Rate Limiting
description: Azure Synapse Analytics has two layers of limits. (1) The management/control plane
  flows through Azure Resource Manager (ARM) and is throttled via the standard ARM token-bucket
  algorithm (250-token read bucket / 25 per sec; 200-token write bucket / 10 per sec per
  subscription per region). (2) The data plane has workload-specific limits — Dedicated SQL
  Pool concurrency by DWU (e.g., DW100c = 4 concurrent queries; DW30000c = 128 concurrent
  queries), Serverless SQL Pool query duration cap, Apache Spark Pool node cap per workspace,
  and Pipelines concurrency caps. Many limits can be raised by support request.
sources:
  - https://learn.microsoft.com/en-us/azure/synapse-analytics/sql-data-warehouse/memory-concurrency-limits
  - https://learn.microsoft.com/en-us/azure/synapse-analytics/sql-data-warehouse/sql-data-warehouse-service-capacity-limits
  - https://learn.microsoft.com/en-us/azure/azure-resource-manager/management/request-limits-and-throttling
headers:
  retryAfter: Retry-After
  remainingSubscriptionReads: x-ms-ratelimit-remaining-subscription-reads
  remainingSubscriptionWrites: x-ms-ratelimit-remaining-subscription-writes
responseCodes:
  throttled: 429
  serviceUnavailable: 503
  resourceGovernance: 10928
limits:
  - name: ARM management plane reads (token bucket)
    scope: subscription/region
    metric: requests_per_second
    limit: 25
    timeFrame: second
    notes: 250-token bucket; 25/sec refill.
  - name: ARM management plane writes (token bucket)
    scope: subscription/region
    metric: requests_per_second
    limit: 10
    timeFrame: second
    notes: 200-token bucket; 10/sec refill.
  - name: Dedicated SQL Pool concurrent queries (DW100c)
    scope: dedicated-sql-pool
    metric: concurrent_queries
    limit: 4
    notes: Concurrency scales with DWU. DW100c=4, DW1000c=32, DW30000c=128.
  - name: Dedicated SQL Pool concurrency slots
    scope: dedicated-sql-pool
    metric: concurrency_slots
    limit: 4
    notes: DW100c=4 slots; DW1000c=40; DW30000c=1200. Each query consumes slots based on resource
      class.
  - name: Serverless SQL Pool query duration
    scope: workspace
    metric: minutes_per_query
    limit: 30
    timeFrame: minute
    notes: Maximum query duration is 30 minutes; long-running queries are auto-killed.
  - name: Apache Spark Pool node count
    scope: workspace
    metric: nodes_per_pool
    limit: 200
    notes: Default cap of 200 nodes per Spark pool; raise via support request.
  - name: Synapse Pipelines concurrent activities
    scope: workspace
    metric: concurrent_activities
    limit: 1000
    notes: Default integration-runtime concurrent-activity cap; raise via support.
policies:
  - name: Pause to control concurrency cost
    description: Dedicated SQL Pools billed by DWU-hour can be paused; metadata is preserved
      and storage continues to bill, but compute charges stop.
  - name: Resource class governance (10928 / 10936)
    description: SQL Pools enforce concurrency via static or dynamic resource classes (smallrc,
      mediumrc, largerc, xlargerc). Exceeding limits returns SQL error 10928.
  - name: ARM token-bucket retry
    description: For management-plane 429s, honor Retry-After and use SDK auto-retry.
  - name: Serverless query result-set size
    description: Serverless SQL Pool returns a result-set up to ~200 GB or 30-minute duration,
      whichever first. Use CETAS / external tables for larger output.
  - name: Limit increase via support
    description: Workspace, pool, and concurrent-activity caps are soft and can be raised via
      Azure support tickets.
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com
    url: https://azure.microsoft.com/services/synapse-analytics/