How should an application handle rate limiting

Hi there,

My understanding is each Algolia plan has a limited queries per second (QPS) – it’s not clear however, what I should expect to happen if I reach this limit – does the API return a 429 and do you recommend a specific backoff algorithm? i.e. with many other third party services, we exponentially backoff requests in an attempt to align with the throttled limit?

Any insight would be much appreciated,
Thanks!