Encountering a persistent 503 Service Unavailable error when the WebRTC softphone attempts to establish the initial WebSocket connection after a fresh environment promotion. The issue is isolated to the genesyscloud_user resource configuration where the softphone_enabled attribute is toggled.
Environment details:
- Genesys Cloud Region: AU (Sydney)
- Terraform Provider: v2.13.1
- Genesys Cloud CLI: v1.10.0
- Browser: Chrome 115 (latest stable)
- Architect Flow: Standard IVR routing, no custom WebRTC logic
The problem manifests immediately after the Terraform apply completes. The user account is correctly provisioned with softphone_enabled = true. However, when the agent logs into the Genesys Cloud Desktop or Web UI, the softphone icon remains grayed out, and the browser console logs show a 503 error on the /api/v2/users/:id/webphone endpoint.
Relevant HCL snippet:
resource "genesyscloud_user" "agent_user" {
first_name = "Test"
last_name = "Agent"
email = "[email protected]"
division_id = var.division_id
softphone_enabled = true
role_ids = [
var.agent_role_id
]
}
Manual verification via the CLI shows the user record is correct:
genesys cloud users get --id <user_id> returns softphoneEnabled: true.
However, the WebRTC signaling server seems to reject the handshake. Restarting the Genesys Cloud Desktop client does not resolve the issue. Only a manual toggle of the softphone setting in the Admin UI (off then on) fixes it temporarily, suggesting a state synchronization lag or a missing dependency in the API call sequence during automated provisioning.
Has anyone observed this 503 specifically on initial provisioning via IaC? Is there a required delay or a specific API call order to ensure the WebRTC license pool is correctly associated before the softphone flag is set? Looking for a programmatic fix rather than manual UI intervention.