Rate limit
Section titled “Rate limit”A rate limit bounds how much work a caller, tenant, or population may request during a defined time interval.
Rate limits protect shared capacity and can enforce fairness between callers. An exceeded limit can delay, reject, or otherwise throttle new work.
A rate limit is an admission policy, not a guarantee that admitted work will finish successfully.
Related knowledge
Load shedding and backpressureKeep overload bounded by slowing producers, limiting admitted work, or rejecting work before saturated components collapse.
Bounded workPlace explicit limits on batch size, concurrency, queue depth, and scan range so one operation cannot consume unbounded resources.