generated: '2026-08-12'
method: searched
source: https://springserve.atlassian.net/wiki/spaces/SSD/pages/1573617663/API+-+Getting+Started
docs:
- https://springserve.atlassian.net/wiki/spaces/SSD/pages/1573617663/API+-+Getting+Started
- https://springserve.atlassian.net/wiki/spaces/CLD/pages/3649929246/CL+-API+-+Getting+Started
applies_to:
- rubicon-project-springserve-v1-api
- rubicon-project-springserve-v0-api
- rubicon-project-clearline-api
limit_count: 3
limits:
- name: Account-wide API cap
  scope: per-account
  limit: 240
  unit: requests
  window: 1m
  endpoints: all
  quote: max 240 request/minute per account
- name: Reporting API cap
  scope: per-account
  limit: 10
  unit: requests
  window: 1m
  endpoints:
  - /api/v1/report
  - /api/v1/reports
  - /api/v0/report
  - /api/v0/reports
  quote: max 10 requests/minute to the reporting API endpoint
- name: Reporting API cap by domain / app bundle
  scope: per-account-per-dimension
  dimension: domain or app bundle
  limit: 3
  unit: requests
  window: 1m
  endpoints:
  - /api/v1/report
  - /api/v0/report
  quote: max 3 requests/minute to the reporting API endpoint by domain/app bundle
response_headers:
  published: []
  observed: []
  note: >-
    Magnite publishes the numbers but no rate-limit response headers. Neither the OpenAPI
    (openapi/rubicon-project-springserve-v1-openapi.yml) nor the getting-started page
    documents X-RateLimit-*, RateLimit-* or Retry-After, and no operation declares a 429
    response — the v1 spec's response codes are 200/201/204/400/401/403/404/422/500. An
    agent therefore has no runtime signal for remaining budget or backoff and must
    self-throttle to the documented ceilings.
exhaustion:
  status_code: null
  note: >-
    Undocumented. No 429 response is declared anywhere in either specification, and the
    behaviour on exceeding the cap is not stated in the docs. Recorded as unknown rather
    than assumed.
retry_guidance:
  published: false
  note: No published backoff or retry policy.
gaps:
- No RateLimit-* / X-RateLimit-* headers published or declared in the spec.
- No 429 (or any) response declared for exhaustion in 668 operations across v1 and v0.
- No documented burst allowance, quota reset semantics, or per-key (as opposed to
  per-account) accounting.