Rate limit calculator
Work out how a token-bucket rate limit behaves for your traffic, and how many requests get throttled.
How it works
A token bucket refills at the limit divided by the window, and holds up to the burst size. Each request takes one token; with none left, the request is throttled.
rate = limit ÷ windowQuestions
What status should a throttled request get?
429 Too Many Requests, with a Retry-After header.
Per client or global?
Per client protects fairness; a global limit protects the service. Many APIs use both.