# Quotas and rate limits

> Rate limits and quotas that apply to every call, and how to back off.

## Rate limits

Write requests are limited to 120 per 60 seconds per client address. Beyond, the answer is `429` with `Retry-After` (seconds): wait that long, then retry with exponential backoff and the same `Idempotency-Key`.

## Service quotas and the free plan

Each service has its own bounds and quotas, read from the code: [Limits and quotas](/cloud/docs/limits). Free plan limits are hard limits: no overage.
