Skip to main content
Requests are metered per API key in a sliding 60-second window. Keys do not compete with each other, so a busy nightly export will not slow down your dashboard.

Exceeding the limit

When a key goes over its window, the API returns a 429 with the standard error envelope. The message tells you how long to wait:
There is no Retry-After header and no X-RateLimit-* headers on the response. Do not parse the wait out of message either: it is prose, and the wording can change. Back off on your own schedule.

Retrying a 429

The window slides, so capacity trickles back continuously rather than resetting all at once. A short first retry usually succeeds. Retry 429 and 5xx responses. Do not retry other 4xx codes: a 400, 401, 403, or 404 means the request itself is wrong, and waiting will not help.
The CLI handles this for you. It retries a 429 up to three times with backoff before surfacing the error.

Spending fewer requests

Errors

The full status-code table, and branching on code.

Pagination

The drain loop, written to spend one request per page.

CLI troubleshooting

What a rate-limited command looks like in the terminal.

Automate with the API

Playbook: a key, the endpoints, and pulling results out.