generated: '2026-08-26' method: probed source: >- Live unauthenticated probes of https://api.mercedes-benz.com on 2026-08-26 (a real 429 was observed and its body and headers recorded), plus the 429 responses declared on all 39 operations in the four provider-published Swagger 2.0 documents in openapi/_original/. Mercedes-Benz publishes no rate-limit reference page that we could read — the developer portal is a client-rendered SPA that returns an empty HTML shell to any non-browser client. evidence: - url: https://api.mercedes-benz.com/remotediagnostic/v1/vehicles status: 429 body: '{ "statusCode": 429, "message": "Rate limit is exceeded. Try again in 58 seconds." }' - url: https://api.mercedes-benz.com/configurator/v1/markets status: 401 note: no RateLimit-* or Retry-After header present on the response limit_count: 0 limit_count_note: >- ZERO published numeric limits. Mercedes-Benz documents that a quota exists — 429 "Quota limit is exceeded" is declared on every operation of every published spec — but publishes no requests-per-second, per-minute or per-day figure for any product or tier. The only runtime number an integrator ever sees is the seconds value inside the 429 message body. apiCommonsRateLimits: "0.1" provider: id: mercedes-me name: Mercedes-Benz Mercedes me url: https://developer.mercedes-benz.com/ description: | Mercedes-Benz documents quota enforcement implicitly through `429 Quota limit is exceeded` responses on every operation in every published Swagger spec (Car Configurator, Dealer, Remote Diagnostic Support, Vehicle Images). Specific request-per-minute and per-day limits are not published in the public spec — they are set per API key/subscription on the developer portal. The Sandbox/Tryout tier is hard-capped on the platform side; the production tiers are governed by partner contract. policies: - id: per-api-key-quota description: Each issued Mercedes-Benz developer API key has a quota counted across all endpoints of the subscribed product. Exceeding the quota returns HTTP 429 with no Retry-After body content per the published Swagger specs (the description string "Quota limit is exceeded" is the only signal). enforcement: per-api-key scope: per-product response: status: 429 bodyDescription: Quota limit is exceeded documented: - https://developer.mercedes-benz.com/ - openapi/mercedes-me-configurator-api-openapi.yml - openapi/mercedes-me-dealer-api-openapi.yml - openapi/mercedes-me-vehicle-images-api-openapi.yml - openapi/mercedes-me-remote-diagnostic-support-api-openapi.yml - id: sandbox-tryout-cap description: The *_tryout basePaths (e.g. /configurator_tryout/v1) are hard-capped for free-tier sandbox use. Exact limit values are dictated by the developer portal subscription, not published in the spec. enforcement: per-api-key scope: per-tryout-basepath response: status: 429 bodyDescription: Quota limit is exceeded - id: customer-consent-throttling description: >- Connected-vehicle APIs (Vehicle Status, Fuel Status, EV Status, Vehicle Lock Status, PAYD) are understood to be throttled per vehicle in addition to per key. enforcement: per-vehicle scope: per-vehicle-per-resource response: status: 429 verified: false evidence: none correction_note: >- DOWNGRADED 2026-08-26. A previous round of this profile stated this policy as fact and gave a reason for it ("to protect on-vehicle modems and avoid waking the vehicle excessively"). Mercedes-Benz publishes no such statement that we could read, and there is no contract for these products to read it from. The policy is retained as an unverified expectation, explicitly marked, rather than deleted or left reading as a measurement. - id: kafka-push-throughput description: >- Fleet API Kafka Push topics are sized per partner subscription. Back-pressure on a streaming channel surfaces as consumer lag or partition rebalancing rather than as an HTTP 429. No public quota number is published. enforcement: per-tenant scope: kafka-topic response: status: n/a bodyDescription: Kafka consumer lag / rebalance verified: false evidence: none correction_note: >- DOWNGRADED 2026-08-26. Retained as an unverified expectation about Kafka semantics in general, not as a measurement of what Mercedes-Benz publishes. Mercedes-Benz publishes no Fleet API throughput figure. observed_live: probed: '2026-08-26' request: 'GET https://api.mercedes-benz.com/remotediagnostic/v1/vehicles (unauthenticated)' status: 429 body: '{ "statusCode": 429, "message": "Rate limit is exceeded. Try again in 58 seconds." }' finding: >- The gateway rate-limits ANONYMOUS requests before it authenticates them, and it returns the retry interval as ENGLISH PROSE inside the message string. An agent must parse a sentence to learn its backoff. This was observed live; it is not in any Mercedes-Benz document. response_headers: retry_after: false ratelimit_limit: false ratelimit_remaining: false ratelimit_reset: false x_ratelimit_family: false checked: >- Response headers inspected on live 401 and 429 responses from https://api.mercedes-benz.com on 2026-08-26. No Retry-After, no RateLimit-*, no X-RateLimit-* header was present on either. consequence: >- This is the single most agent-hostile property of the estate. Every published operation declares a 429, and none of them gives a machine a number. A well-behaved client cannot compute a backoff from headers; it must regex an English sentence out of a JSON message field. notes: - All four published Swagger specs explicitly document `429 Quota limit is exceeded` as a response on every operation, demonstrating consistent platform-wide quota enforcement. - Exact RPS / RPM / requests-per-day numbers are quoted per contract and not in the public OpenAPI specs. - There is no public Retry-After contract documented; clients should treat 429 as exponential-backoff.