When an error happens the Runa API returns a response with a status code of 4xx or 5xx and a JSON body with details of the error. Our error bodies follow this format: 
{
  "type": "<the error type>",
  "message": "<a human readable message describing the error>",
  "help": "<additional information to help resolve the error>"
}
Some specific errors include additional keys, for example validation errors, to help pin down the issue. Check the reference page for an endpoint to see the specific HTTP status codes that are expected. You can use the status code to determine if that error provides additional keys. For example the 422 response on the create order endpoint includes a details key to identify which fields are invalid.

​
Error types

Error type reference

The reference section page on error types lists all possible errors you might encounter when using the Runa API.
 You can use the type field to determine the type of error that occurred and take appropriate action. You must not rely on the message or help fields as these may change over time.

​
Handling final vs transient errors

When an error occurs, it’s important to determine whether it’s a final error or a transient error. A final error is an error that will not be resolved by retrying the request. A transient error is an error that may be resolved by retrying the request.

​
Final errors

Final errors have 4xx status codes, indicating an issue with the request being made. For example an invalid payload or a specific error with the payload such as an unsupported value for a product. In these cases you should correct the issue and retry the request using a new idempotency key. Using the same idempotency key will result in the same error being returned.

​
Transient errors

Transient errors have 5xx status codes, indicating an issue with the Runa API. These errors are usually temporary and can be resolved by retrying the request. In these cases you should retry the request using the same idempotency key. If the error persists you should wait a short period of time before retrying the request.