# Rate limits

How many requests a key may send, and what to do when it sends too many.

| Limit | Requests a minute |
| --- | --- |
| Per credential (key or token) | 600 |
| Per organization, across all its credentials | 1,200 |

A request over either limit gets **429** `request.rateLimited` with a `Retry-After` header: the number of seconds to wait before the next request.

## Handling a 429

<Steps>

<Step title="Wait the time you were given">

Read `Retry-After` and pause that many seconds. Don't retry in a tight loop; every refused request still counts.

</Step>

<Step title="Spread the work">

Batch reads with `limit=100` instead of many small pages, and run background syncs off-peak.

</Step>

<Step title="Split integrations">

Give each integration its own key, so one busy sync can't use up another's minute. They still share the organization's limit.

</Step>

</Steps>