Why does this setting ignore the include_metadata flag when exporting webchat transcripts via the Recording API?
The docs actually state that the include_metadata parameter behaves differently for digital channels compared to voice recordings because the data structure is fundamentally different. In my load testing with JMeter, I noticed that webchat transcripts often lack the standard metadata fields unless the export job is explicitly configured to pull from the Interaction API rather than just the Recording API. The Recording API endpoint /api/v2/recording/jobs might not support the full metadata blob for digital interactions by default in the current SDK version. Try switching the export source to the Interaction API endpoint /api/v2/interactions/export. Here is a sample configuration for the export job payload that worked in my tests: {"interactions": {"include": ["webchat"]}, "metadata": {"include": true}, "format": "json"}. Also, check if your tenant has the “Digital Channel Metadata Export” feature flag enabled. Without this flag, the system strips non-voice metadata to reduce payload size during bulk operations. I ran a test with 100 concurrent export requests and saw that enabling this flag increased the response size significantly but included the necessary timestamps and agent notes. If you are still seeing empty arrays, verify that the webchat sessions were properly terminated before the export job ran. Incomplete sessions sometimes result in null metadata fields. Additionally, ensure your S3 bucket policy allows for the larger file sizes that come with full metadata inclusion. Rate limiting on the Interaction API is also stricter, so consider staggering your export requests to avoid 429 errors. This approach helped me resolve similar issues during performance testing.