generated: '2026-08-13'
method: searched
source: https://docs.adverity.com/guides/management-api/introduction-management-api.html
docs: https://docs.adverity.com/guides/management-api/introduction-management-api.html
limit_count: 1
policy: none-published-except-one-route
statement: >-
  "Adverity's Management API has no rate limits. However, each Management API call uses
  resources, so we recommend that you only perform one Management API task at a time. This
  avoids overloading your resources and will prevent performance issues. Wait for one
  Management API call to finish before performing another task."
limits:
- scope: per-ip
  endpoint: GET /api/v1/auth/keys/introspect/
  window: not-published
  limit: not-published
  burst: null
  reason: >-
    IP-based rate limiting was implemented on the unauthenticated API-key introspection
    route to prevent denial-of-service attacks (release 2026.30). A follow-up fix in the
    same release stopped the introspection rate limit from throttling multiple tenants at
    once, specifically to avoid unintended throttling of requests reaching MAPI through the
    MCP server.
  source: https://docs.adverity.com/reference/release-notes/release-notes.html
  note: >-
    The only rate limit Adverity acknowledges anywhere. It applies to the one route that can
    be called without a key. Neither the numeric limit nor the window is published.
headers:
  documented: false
  observed: []
  note: >-
    No X-RateLimit-*, RateLimit-*, or Retry-After response headers are documented, and none
    could be observed: every Management API host is per-instance
    (https://{{INSTANCE}}/api/) and requires an authenticated key, so there is no
    anonymously reachable endpoint to read live headers from. The MCP endpoint
    (https://mcp.eu.adverity.com/mcp) returned 401 with only a WWW-Authenticate header.
exhaustion:
  status_code: null
  note: Not documented — Adverity states there are no rate limits to exhaust.
guidance:
  concurrency: >-
    The documented control is serialization, not a quota: run one Management API task at a
    time and wait for it to finish. This is advisory, has no machine-readable signal behind
    it, and is enforced by resource contention on the customer's own instance rather than
    by the API.
downstream_limits:
  note: >-
    Adverity does surface rate limiting of the SOURCE platforms it collects from — for
    example release 2026.33 records that the TikTok Ads connector was improved to handle
    QPS limits more effectively. Those are third-party limits, not Adverity API limits.
note: >-
  Effectively an honest zero. Adverity affirmatively documents that its Management API has
  no rate limits — a published position rather than an absence of documentation — and the
  single exception is an IP limit on the one unauthenticated route. Either way an agent gets
  no runtime backpressure signal (no headers, no documented 429 contract) and must
  self-throttle on the "one task at a time" advice.
related:
- conventions/adverity-conventions.yml