Questions about Rate Limiting come up in regular technical interviews as well as in System Design interviews.
Imagine you are building an API that is available to all registered clients. Several potential problems arise right away.
First, if users can make an unlimited number of requests, then as the product grows this will significantly increase the load on the system and infrastructure costs.
It will also lead to an uneven distribution of resources between clients and to the Noisy Tenants problem (when one client consumes a large share of the resources, which in turn negatively affects the others).
In this case, you can limit the number of API calls. For example, clients on the basic plan can be allowed to make no more than 1000 requests per day. If they need more, they can upgrade to a different plan with higher limits.
When a client exceeds the set limit, the server usually returns the status 429 Too Many Requests. Additionally, the response can indicate how long the client needs to wait before the next attempt.
Thus, Rate Limiting helps control load, reduce costs, and ensure that system resources are distributed fairly among all clients.