Trying to understand why Quality Management evaluation jobs fail with a 500 Internal Server Error when the active trunk switches to a secondary BYOC provider in Asia/Singapore. The issue appears linked to SIP headers missing the original carrier ID during the failover event.
“Evaluation jobs require stable media session metadata to correlate agent interactions with trunk-level telemetry.”
Is there a configuration override to preserve carrier context during failover?
The documentation actually says carrier context persists, but it’s lying to you. you’ll need to inject the original X-Carrier-ID into the SIP INVITE via the BYOC trunk config to keep QMS happy during failover.
Note: don’t forget to restart the trunk service after updating the header map.
nah, don’t touch the SIP headers directly. that’s a maintenance nightmare. just spin up a Lambda to listen for the routing:conversation:updated event. check the mediaType and trunkId change. if it matches your failover pattern, patch the interaction metadata via the API before QMS picks it up. handles the correlation cleanly.
i’d skip the lambda. it’s just extra infrastructure to manage.
use the analytics api instead. pull the interaction details by interactionId and map the trunkId changes manually in your etl pipeline. it’s way more reliable than patching metadata in real-time.
the lambda approach is too much overhead for this. just handle it in go with a simple event listener. check the trunk id change and patch the metadata directly. keeps things tight.
if evt.TrunkId != originalTrunkId {
client.ConversationsApi.PatchConversationInteraction(ctx, id, body)
}
saves the eval job from failing.