Dealing with a very strange bug here with the bulk export jobs when trying to pull predictive routing data for digital channels. We are operating in a BYOC environment on EU-West-1. The goal is to export interaction metadata specifically for legal discovery requests involving WhatsApp and MMS interactions that were handled via predictive routing queues.
Standard media file exports work fine. The issue arises when we attempt to include routing metadata in the export payload. The job fails immediately with a 422 Unprocessable Entity error. The error message is vague, just stating ‘Invalid filter criteria for requested resource type’.
Here are the steps to reproduce the issue:
- Create a new bulk export job using the Bulk Export API v2.
- Set the resource type to
INTERACTION. - Apply a filter for
channelequal towhatsappormms. - Add a filter for
routing_queue_idto target specific predictive routing queues. - Include
routing_metadatain the requested fields array. - Submit the job.
The job status changes to FAILED within seconds. The logs show no detailed reason other than the 422 status. I have verified that the queue IDs are correct and that interactions exist for the specified time range. Removing the routing_metadata field allows the job to complete successfully, but this leaves us without the necessary audit trail for chain of custody purposes.
We are using the standard S3 integration for destination. The bucket permissions are correct, as other export jobs write to it without issue. This seems to be specific to the combination of digital channels and predictive routing metadata.
Has anyone successfully exported predictive routing metadata for digital channels in a BYOC setup? Is there a known limitation or a specific field format required for the routing metadata request? Any insights would be appreciated. We need this data for an active legal case, so time is of the essence.