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: Google Marketing Platform Admin
providerId: google-marketing-platform
generated: '2026-08-13'
method: searched
source: https://developers.google.com/marketing-platform/devguides/api/admin/v1/quotas
created: '2026-05-04'
modified: '2026-08-13'
tags:
  - Analytics
  - Google Marketing Platform
  - Marketing
  - Organization Management
  - Platform Administration
  - Rate Limiting
  - Quotas
  - Throttling
description: >-
  Published rate limits for the Google Marketing Platform Admin API, read from
  the provider's own "Limits and Quotas on API Requests" page on 2026-08-13.
  This file REPLACES a 2026-05-04 bulk-sweep scaffold that carried invented
  tier limits (10/100/1000 rpm, X-RateLimit-* headers) which the provider does
  not publish and does not return. Quotas are enforced per Google Cloud
  project, refilled automatically every 60 seconds, and are visible per project
  in the Cloud console at
  https://console.cloud.google.com/apis/api/marketingplatformadmin.googleapis.com/quotas
limit_count: 4
headers:
  note: >-
    The provider documents no rate-limit response headers for this API. Google
    API exhaustion is signalled in the error body, not in RateLimit-* or
    X-RateLimit-* headers, and no Retry-After is documented. An agent must read
    error.errors[].reason == rateLimitExceeded from the JSON error envelope
    rather than a header. Recorded as an honest absence, not a guess.
  limit: null
  remaining: null
  reset: null
  retryAfter: null
  policy: null
responseCodes:
  throttled: 403
  throttled_reason: rateLimitExceeded
  note: >-
    Google returns HTTP 403 with reason rateLimitExceeded on quota exhaustion
    for this API, not the more common 429. This is documented explicitly on the
    quotas page and is a real interoperability trap for agents that treat 403
    as a permanent authorization failure and stop retrying.
limits:
  - name: Requests per minute
    scope: project
    metric: requests_per_minute
    limit: 1200
    timeFrame: minute
    window_seconds: 60
    applies:
      - Google Marketing Platform Admin API
    note: Consumed by every request, read or write.
  - name: Requests per minute per user
    scope: project-per-user
    metric: requests_per_minute
    limit: 600
    timeFrame: minute
    window_seconds: 60
    applies:
      - Google Marketing Platform Admin API
    note: Consumed by every request, read or write, counted per end user.
  - name: Writes per minute
    scope: project
    metric: writes_per_minute
    limit: 300
    timeFrame: minute
    window_seconds: 60
    applies:
      - Google Marketing Platform Admin API
    note: >-
      Consumed in addition to the requests quotas by create, patch, delete,
      archive and update methods. For this API that is
      createAnalyticsAccountLink, deleteAnalyticsAccountLink and
      setPropertyServiceLevel.
  - name: Writes per minute per user
    scope: project-per-user
    metric: writes_per_minute
    limit: 120
    timeFrame: minute
    window_seconds: 60
    applies:
      - Google Marketing Platform Admin API
policies:
  - name: Refill interval
    description: >-
      Rate limits are enforced and automatically refilled in 60-second
      intervals. On rateLimitExceeded the documented guidance is to wait a
      minute and retry; the quota is refilled at the start of the next
      interval.
  - name: No daily quota
    description: >-
      The provider states there is currently no explicit daily usage quota for
      this API; the maximum number of daily requests is naturally bounded by
      the per-minute limits.
  - name: Per-user capping
    description: >-
      Per-user limits can be managed through the Google Cloud "Capping API
      usage" feature.
    url: https://cloud.google.com/apis/docs/capping-api-usage
  - name: gRPC client retry policy
    description: >-
      The first-party gRPC service config published with the protos sets a
      60s timeout and a retry policy of maxAttempts 5, initialBackoff 1s,
      maxBackoff 60s, backoffMultiplier 1.3, retrying only UNAVAILABLE and
      UNKNOWN. RESOURCE_EXHAUSTED / rateLimitExceeded is NOT in the retryable
      set, so the generated clients will not retry a quota error for you.
    source: grpc/google-marketing-platform-grpc-service-config.json
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com