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: Supaglue
providerId: supaglue
created: '2026-05-04'
method: searched
modified: '2026-08-13'
generated: '2026-08-13'
source: >-
  https://github.com/supaglue-labs/supaglue/blob/main/docs/docs/api/introduction.mdx and
  openapi/_original/supaglue-management-api-openapi.yml
reconciled: false
limit_count: 0
tags:
  - CRM
  - Unified API
  - Open Source
  - Rate Limiting
description: >-
  Supaglue published no numeric rate limit for its own API — not on the API introduction page, not in
  any spec, and not in the archived docs anywhere. What it published is the exhaustion signal (HTTP
  429, error code TOO_MANY_REQUESTS_ERROR, in the error table on the API introduction page) and,
  notably, an operation that reports the THIRD-PARTY provider's remaining quota rather than its own:
  GET /customers/{customer_id}/connections/{provider_name}/_rate_limit_info. That inversion is the
  finding — Supaglue treated the upstream SaaS limit as the limit that mattered, because a unified
  call is really a Salesforce or HubSpot call wearing a different URL. No X-RateLimit-* or
  RateLimit-* response headers and no Retry-After are documented, so a client cannot see how close it
  is to a wall until it hits one. The hosted service is now shut down (api.supaglue.io does not
  resolve); on a self-hosted deployment the throttle is whatever the deployer configures.
sources:
  - https://github.com/supaglue-labs/supaglue/blob/main/docs/docs/api/introduction.mdx
  - https://github.com/supaglue-labs/supaglue/blob/main/openapi/v2/mgmt/openapi.bundle.json
notes: >-
  limit_count is 0 because zero numeric limits were ever published for Supaglue's own API — an honest
  zero, searched for and not found, not an unmeasured gap. Probed 2026-08-13.
responseCodes:
  throttled: 429
  throttledCode: TOO_MANY_REQUESTS_ERROR
  remoteProviderError: 499
responseHeaders:
  published: []
  note: No rate-limit response headers are documented or present in any spec.
limits: []
introspection:
  - name: Connection rate limit info
    operationId: getConnectionRateLimitInfo
    path: GET /customers/{customer_id}/connections/{provider_name}/_rate_limit_info
    source: openapi/_original/supaglue-management-api-openapi.yml
    reports: the third-party provider's remaining quota for that customer's connection
    note: This is the only rate-limit visibility in the product, and it is about the upstream SaaS.
policies:
  - name: Pass-Through Throttling
    description: >-
      A unified request is executed against the third-party provider unless read_from_cache=true, so
      the effective ceiling is the provider's own limit — Salesforce, HubSpot, Pipedrive and the rest
      each govern their own. Reading from the Managed Destination cache is the documented way to
      avoid spending provider quota.
  - name: Self-Hosted Deployment Limits
    description: >-
      With the hosted service gone, any throttling is defined by whoever runs the MIT-licensed code.
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com