> ## Documentation Index
> Fetch the complete documentation index at: https://docs.junction.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Rate Limiting

> Learn about Junction API rate limiting behavior, HTTP 429 and 503 responses under server load, and targeted limits on specific endpoints.

Junction does not impose a per-customer rate limit on server-to-server API calls at this time. We adjust the
elastic capacity and overprovisioning rate of our API servers regularly, in response to your usage pattern changes
and product launch plans.

However, when Junction API servers are under stress, there is a possibility that our infrastructure may abort a number
of your API requests with a **429 Too Many Requests** or **503 Service Unavailable** response.

For idempotent API requests, consider retrying the API request at least once, upon receiving a 429 or 503 response. Optionally,
if you are making the API request in an asynchronous job context, consider having more retry attempts with an
exponential backoff strategy.

## Targeted rate limits

You may encounter **429 Too Many Requests** in these specific scenarios:

| Endpoint                       | Authentication Scheme | Rate Limit          |
| ------------------------------ | --------------------- | ------------------- |
| `POST /user/refresh/{user_id}` | Junction API Key      | 8 per hour per user |

<Note>
  Junction may adjust rate limiting based on real-world usage patterns. We will give you sufficient notice
  if a new enforcement would impact your existing usage patterns.
</Note>