Stumbled on a weird bug today with our custom data action that calls an external CRM API. it works fine 99% of the time but fails hard when we trigger a manual failover on our BYOC trunks in the AP-SE-2 region.
the flow is simple: inbound call → gather input → data action → update CRM. when the trunk switches from primary carrier to secondary, the data action step times out or returns a 502 Bad Gateway. the error log in architect shows “connection reset by peer”.
we are using the latest version of the platform. the external API is hosted in AWS ap-southeast-1. latency is usually under 200ms. during failover, the call stays connected but the data action hangs for 15 seconds then fails.
could this be related to how genesys handles outbound connections from the edge node during trunk re-registration? we have 15 trunks configured. the failover logic seems to interrupt the data action execution context.
anyone else seen this? is there a retry mechanism we can add in the flow or is this a platform limitation? the error is consistent across all 15 trunks when failover is triggered. not sure if it is a DNS issue or something with the edge node routing. please advise.