Connection terminated with error code ICE_FAILED after 30 seconds of active conversation. This is happening on the standard web softphone interface within the EU-FR environment. The Architect flow is using a standard Make Outbound Call block with no custom logic interfering with the signaling. We have verified that the network allows UDP traffic on ports 10000 through 59999 as per the Genesys Cloud documentation for BYOC setups. Despite this, the WebRTC negotiation fails intermittently. The performance dashboard shows the agent as “Available” but the call never rings on the destination side. There are no 5xx errors in the general system status. Is there a specific configuration in the EU-FR edge nodes that requires additional firewall rules for WebRTC? The issue started after the last platform update on 2023-10-15. We need to understand if this is a regional latency issue or a configuration mismatch in the softphone settings. The impact on our agent productivity is significant as manual dialing is the only workaround currently.
{
"ice_servers": [
{"urls": "stun:stun.l.google.com:19302"},
{"urls": "turn:eu-fr-turn.genesys.cloud:443", "credential": "key", "username": "user"}
]
}
You might want to look at the ICE server configuration in your softphone settings. The standard Google STUN server often fails in EU-FR due to local carrier NAT traversal issues, so forcing a Genesys-specific TURN server usually resolves the ICE_FAILED drop.
terraform plan shows no diff on the ice_servers block after adding the eu-fr-turn config. deployments are now stable across the eu-fr region. thanks for the pointer.