...
The batch does not abort on failure but keeps on processing the subsequent records, unless at least 20% of the total failed. In the latter case, the batch is assigned the status TooManyFailed. Failed records can be retrieved through monitoring (Batches → Failed records) or the API: /batches/:batchId/failures
API Permissions
POST batches/
...