Has anyone else found this issue ?
We only looked into it when we had a lambda that normally returns within 5 seconds started to take 60 seconds due to other back end issues.
Strangely we saw after 30 seconds the data action refired. We have CloudTrail Invoke logs proving that the source of the re-fire was Genesys Cloud. It's with Genesys now as a case.
No error is returned by the lambda because it completes successfully after 60 seconds (a simple test with lambda containing one line of code to sleep for 45 seconds can trigger the issue).
I am hoping to discover if anyone else had experienced this and how you decided to work around the issue.
The Data Action verb is already POST.
When we created the test lambda and data action we also replicated the HTTP Method to simulate the same conditions. Since we can replicate the issue and capture the CloudTrail invoke logs we were able to prove the origination of the second request.
Even if Genesys discover the reason for it and eventually fix it, we'll likely still have to implement a workaround in the lambda to accommodate the unwanted duplication in the interim.