Hey everyone, I’ve run into a really strange issue with the bulk export process for WebRTC softphone interactions. We are pulling data for a legal discovery request that requires strict chain-of-custody documentation. The issue is that while the audio files download correctly to our S3 bucket, the associated metadata JSON files are missing critical fields for interactions initiated via the WebRTC softphone.
The specific fields missing are recording_start_time and agent_screen_share_url. These fields are present for standard SIP trunk recordings but return null for WebRTC sessions. We are using the POST /api/v2/bulkexports/jobs endpoint with a filter object targeting recording_type: "screen_and_audio" and interaction_type: "webchat" (as the softphone session is linked to a webchat context in our flow).
The job completes with a status of completed, but the resulting manifest indicates missing data for approximately 15% of the records. All affected records share a common trait: they were handled by agents using the Genesys Cloud WebRTC softphone in Chrome 124.0.6367.159. The Architect flow logs show the recording command was executed correctly with record: true set on the action.
We have verified that the legal hold is active on the interaction, which should force the retention of all metadata. However, the export job seems to ignore the WebRTC-specific metadata structure. Is there a known limitation with the bulk export API regarding WebRTC softphone metadata? We need to ensure the audit trail is complete for the legal team. Any insights on how to force the inclusion of these fields or if we need to query the analytics API directly for these specific timestamps?