Breaking change on 18-11-2026: new error format for quota and rate limits
9 days ago by Company.info
On 18-11-2026, the error responses for quota and rate limits change to the JSON:API error format. Each error gets a fixed code and a link to the documentation.
This is a breaking change if your integration reads the body or the Content-Type of these errors. If it does, update your error handling.
What changes
Since 10-11-2026, two changes are already live:
429and503responses include aRetry-Afterheader with the number of seconds to wait.- When the service is busy, the API returns
503 Service Unavailableinstead of429 Too Many Requests.
On 18-11-2026, the body and the Content-Type of these responses change as well.
| Response | Meaning | Old detail | New code |
|---|---|---|---|
403 | Quota used up | Quota exceeded | quota_exceeded |
429 | Rate limit exceeded | Rate Limit Exceeded | rate_limit_exceeded |
503 | Service busy | API Rate Limit Exceeded | capacity_exceeded |
Other errors do not change.
Before 18-11-2026
HTTP/1.1 429 Too Many Requests
Content-Type: application/json
Retry-After: 2{
"type": "about:blank",
"title": "Too Many Requests",
"detail": "Rate Limit Exceeded",
"status": 429
}From 18-11-2026
HTTP/1.1 429 Too Many Requests
Content-Type: application/vnd.api+json
Retry-After: 2{
"errors": [
{
"status": "429",
"code": "rate_limit_exceeded",
"title": "Rate limit exceeded",
"detail": "Too many requests per second ...",
"links": { "type": "..." }
}
]
}The fields move into the errors array, and status is now a string. The 403 and 503 responses change in the same way.
Update your integration
- Retry a
503like a429. Wait for the number of seconds inRetry-After, then retry. - Read the new format. Use
errors[0].codeto find the cause. Do not match ontitleordetail. These texts are for people and can change. - Handle the new
Content-Type:application/vnd.api+json. - Check for
quota_exceededon a403. Do not retry it automatically: wait until the quota renews. A403without this code means something else, usually that your key has no access to this API.
For how to handle each response, see Quota & Rate Limits.
