Quick question about WebRTC softphone configuration drift in AU-1 BYOC

Voice & Telephony
1

Quick question about managing WebRTC softphone settings via CX as Code pipelines. We are attempting to standardize the genesyscloud_webrtc resource across our dev and prod environments in AU-1 BYOC. The Terraform plan shows no drift, but the actual runtime behavior differs. Agents in Prod report audio latency spikes that are not present in Dev.

Environment:

  • Genesys Cloud CLI v2.1.0
  • Terraform Provider v1.42.0
  • Region: AU-1 BYOC
  • GitHub Actions runner: ubuntu-latest

We are using the following HCL to configure the WebRTC resource:

resource "genesyscloud_webrtc" "main" {
 enabled = true
 display_name = "Corporate WebRTC"
 
 webrtc_config {
 ice_servers {
 urls = ["stuns:stun1.genesiscloud.com:443", "stuns:stun2.genesiscloud.com:443"]
 username = null
 credential = null
 }
 
 constraints {
 audio {
 echo_cancellation = true
 noise_suppression = true
 auto_gain_control = true
 }
 }
 }
}

The CLI output confirms the resource is applied successfully:

$ genesyscloud webrtc list
Id: abc-123-def
Name: Corporate WebRTC
Enabled: true

However, when we query the API directly via /api/v2/users/{userId}/webrtc/status, the iceServers array in Prod contains additional turn servers that are not defined in our Terraform state. This suggests that the platform is overriding our configuration or that there is a caching issue in the BYOC tenant.

Reference: WebRTC Configuration Docs

Has anyone seen this behavior where the API returns different ICE servers than what is provisioned via CLI/Terraform? We need to ensure deterministic behavior for our automated audits. Any insights on how to force a refresh or verify the actual active configuration would be appreciated.