name: Atlan API Rate Limits
description: >-
  Atlan's REST API enforces rate limiting on a per-tenant, per-token basis.
  The platform returns HTTP 429 (Too Many Requests) when limits are exceeded.
  Atlan's official SDKs implement automatic retry logic with exponential backoff
  to handle transient throttling transparently. Specific numeric thresholds are
  not publicly documented and vary by deployment tier and tenant configuration.
specificationVersion: "0.1"
url: https://docs.atlan.com/
rateLimits:
  - name: REST API Throttling
    description: >-
      Atlan's REST API enforces rate limits per API token or OAuth client.
      When the limit is exceeded the API returns HTTP 429. The Python SDK
      retries up to 5 times with exponential backoff by default; the Java SDK
      retries up to 3 times. Customers executing 10,000+ API calls per week
      are documented in Atlan's own case studies.
    statusCode: 429
    retryAfterHeader: Retry-After
    sdkRetryBehavior:
      python:
        maxRetries: 5
        backoff: exponential
        retriableStatusCodes:
          - 403
          - 429
          - 500
          - 502
          - 503
          - 504
      java:
        maxRetries: 3
        backoff: exponential
        retriableStatusCodes:
          - 403
          - 429
          - 500
          - 502
          - 503
          - 504
    notes: >-
      Specific rate-limit thresholds (requests per minute/hour) are tenant-specific
      and not publicly disclosed. Contact Atlan support or your customer success
      manager for limits applicable to your deployment.
  - name: API Token Scope
    description: >-
      Each API token carries the full permissions of the issuing user or service
      account, filtered by the persona assigned to the token. OAuth 2.0 clients
      use scoped, short-lived tokens for least-privilege access.
    authType: BearerToken
    alternativeAuth: OAuth2ClientCredentials
  - name: Search / Query Limits
    description: >-
      Search result sets are paginated. Large bulk operations (asset creation,
      bulk classification) should be batched and throttled by the caller. The
      Atlan SDKs provide iterator helpers for safe pagination of large result sets.
    pagination: cursor-based
    notes: >-
      Bulk operations on very large asset sets may require partitioning workloads
      across time windows. SDK batch helpers are available in Python and Java.
documentation:
  url: https://docs.atlan.com/get-started/references/api-authentication
  support: https://ask.atlan.com/