The bulk export job for SIP trunk recordings is returning a 200 OK status, but the downloaded JSON metadata files are missing the call_id field for approximately 15% of the interactions. The error log in the Genesys Cloud admin console shows ExportItemError: MetadataIncomplete for these specific entries. This is causing a chain of custody gap for our legal discovery requests, as we cannot map the audio blobs back to the specific SIP dialog without this unique identifier.
We are running this export against the eu-west-1 region using the /api/v2/recording/search endpoint with a date range filter. The interactions in question are inbound calls routed via a SIP trunk connected to a third-party telephony provider. The Architect flow uses a standard Record Call action with the record_on_hold setting enabled. We have verified that the recordings themselves are present in the S3 bucket and are playable, but the metadata drift issue persists. This seems to happen only for calls that are transferred after the initial recording segment starts, suggesting a potential issue with how the system handles metadata updates during transfer events.
Is this a known issue with the current version of the Recording API for SIP trunk interactions? We need to ensure complete metadata integrity for compliance. Any advice on how to manually reconstruct the call_id mapping or force a metadata refresh for these affected recordings would be greatly appreciated. We are currently stuck on holding these records until the data is reconciled, which is impacting our SLA for legal requests.