WFM Schedule API 429 errors during peak shift swap

Stumbled on a weird bug today with the workforce management schedule endpoints. specifically hitting 429 too many requests errors when querying agent availability for bot fallback logic. the integration hits these 429s consistently during UK business hours des… wait no that was last week. today it is different. today the issue is with the data actions connector for servicenow. when the architect flow triggers a screen pop via the webchannel api and simultaneously tries to create a ticket in servicenow via the data actions outbound webhook, the servicenow call times out with a 504 gateway timeout. but only when the agent is in a wrap-up state. if the agent is available, the ticket creates fine. if the agent is busy, the ticket creates fine. but wrap-up state causes the webhook payload to be malformed or dropped by the genesys cloud data actions service. checking the architect logs shows the data action step completes with a success status but the servicenow side never receives the payload. i have checked the servicenow rest api logs and there is no request recorded. this suggests the issue is within the genesys cloud data actions service dropping the request due to some internal rate limiting or queue saturation related to the wfm state changes. the environment is genesys cloud 2024-08 release. using the standard data actions servicenow connector. the payload is a simple json object with incident short description and caller id. no custom headers. has anyone seen this specific behavior where wfm state affects data actions reliability? i have tried adding a wait step before the data action call but it does not help. the error is inconsistent. sometimes it works during wrap-up, sometimes it fails. looking at the network traces on the browser side, the webchannel messages are being sent correctly. the issue seems to be server-side within the data actions execution engine. any insights on how to debug this further? should i be using a different approach for creating tickets during wrap-up? maybe a scheduled job instead of real-time webhook? or is there a known issue with the data actions connector and wfm state transitions? please advise.