Trying to understand why our bulk export jobs for digital channel recordings are timing out when we include full metadata for legal discovery purposes. We are using the /api/v2/bulkexport/jobs endpoint to initiate exports for WhatsApp and Web Chat sessions from the last quarter. The requirement from our legal team is strict: every message must have the complete chain of custody metadata, including agent IDs, queue names, and interaction timestamps.
The job starts successfully, returning a 202 Accepted status. However, after approximately 45 minutes, the job status flips to failed without a specific error code in the initial response. When we poll the job details via /api/v2/bulkexport/jobs/{jobId}, the error field is empty, but the download URL is null. We have checked the S3 bucket permissions, and the IAM role has full write access. The issue seems specific to the size of the metadata payload. When we run the same export without the metadata flag, the job completes in under 10 minutes.
We are operating in the EU-West region, and our API client is using the latest Python SDK version 2.14.0. The digital channel volume is around 50,000 interactions per day. We suspect the API might be hitting an internal processing limit for metadata serialization on digital channels, which have more complex structures than voice recordings. Has anyone seen this behavior with large-scale digital channel exports? We need a reliable way to get this data for a pending discovery request by Friday. Any insights on whether this is a known limitation or if there is a workaround, such as chunking the date ranges further? We have already tried reducing the date range to 7 days, but the timeout persists if metadata is included.
Thanks for the help.