Dealing with a very strange bug here with the screen recording export pipeline when processing high-volume legal discovery requests. We are pulling data via the POST /api/v2/recordings/screen/bulkexport endpoint to fulfill chain of custody requirements for a specific client. The environment is Genesys Cloud version 2024-03.01.01, and we are using the Python SDK version 3.1.4 for the orchestration script.
The issue is not with the audio or video files themselves; those download correctly to the S3 bucket. The problem lies in the associated metadata JSON file generated for the bulk job. Specifically, the start_time and end_time fields in the metadata are consistently offset by 30 minutes compared to the actual interaction timestamps in the GC portal. This discrepancy breaks our audit trail validation scripts, which require exact timestamp matching to verify recording integrity.
Error: Metadata timestamp mismatch detected in export manifest
We have verified that the user account running the export has the correct timezone settings (Europe/London, GMT+0/BST+1) and that the screen recording configuration in the admin panel is set to capture based on interaction time, not agent clock time. We tried adjusting the date_range parameters in the API call to explicitly define UTC boundaries, but the output remains inconsistent.
Has anyone else encountered this specific timestamp drift in screen recording bulk exports? Is there a known issue with how the screen recording service syncs metadata with the main interaction database during peak load periods? We need to resolve this urgently as it impacts our ability to certify the recordings for legal submission. Any insights into whether this is a configuration error on our end or a platform-side bug would be greatly appreciated.