What it means
Multiple limits can apply
Services can impose global, endpoint-specific, resource-specific or concurrency limits. These are different controls: reducing the number of requests per second may not solve a limit on simultaneous operations.
Responding without duplication
HTTP 429 indicates that a client is sending too many requests under the applicable policy; a response may include retry guidance. Use the provider’s documented backoff and recovery behavior. For writes, first preserve operation identity so a delayed retry does not accidentally create an additional payment.
Related concepts
Application programming interface, Idempotency, Service-level agreement.
In practice
A concrete example
An integration reduces concurrency and schedules retries with jitter while retaining the original operation’s idempotency identifier.
Technical notes & boundaries
Not every 429 response has the same cause, and Retry-After is not mandatory in every response. Limits are provider- and account-specific.
Go to the source.
Original references used for this definition. Product details and standards can change.