Could someone explain the expected behavior of SIP trunk registration stability when the underlying OAuth token used for authentication expires during high-volume call processing? We are deploying a custom telephony integration via AppFoundry that bridges external SIP trunks into Genesys Cloud using the platform API to manage trunk configurations. The application runs on AWS in us-west-2, matching the PSTN connectivity region, but we observe intermittent 401 Unauthorized responses on OPTIONS keepalive requests exactly when the access token refreshes. The SDK version is @genesyscloud/genesyscloud-pstn-trunking@2.4.1, and the flow logic relies on POST /api/v2/architect/flows to route calls, which fails if the trunk drops. The error logs show SIP/2.0 401 Unauthorized with a WWW-Authenticate header that seems to ignore the cached credentials. We have verified that the refresh grant flow is functioning correctly for other API endpoints, but the SIP signaling layer appears to lose context. Is there a specific header or retry mechanism required in the SIP INVITE or REGISTER request to handle this token rotation without dropping active calls?