I can’t seem to figure out why the terraform import command consistently fails to reconcile the existing notification_webhook resource with my local state file. The documentation explicitly states, “When importing resources, ensure the provider configuration matches the target environment’s API version exactly.” I have verified that my provider block targets version 1.45.0, yet the state synchronization aborts immediately after the handshake phase. I executed the import directive against the production tenant endpoint, and the CLI returns a malformed reference error on line 14 of the generated tfstate json. To isolate the variable mapping issue, I manually constructed and sent the following payload to the configuration validation endpoint before triggering the import routine.
{
"uri": "/api/v2/notification/webhooks/8f3a1c92-4b7d-4e11-9c8a-22f1a6b3c001",
"name": "Supervisor WebSocket Sync Hook",
"enabled": true,
"targets": [
{
"uri": "https://dashboard.internal/events",
"type": "webhook"
}
],
"subscriptions": [
{
"topic": "routing.queue",
"event_types": ["routing.queue.stats"]
}
]
}
The response code is 200, which confirms the resource exists and matches my schema definitions. However, when I pipe this structure into the state import workflow, the provider throws a 409 conflict during the reconciliation step. I have disabled the automatic retry logic in my provider configuration to capture the raw response headers. The timestamp formatting in the state file appears correct for the Asia/Tokyo timezone, yet the diff calculator treats the subscription array as a completely new block rather than an in-place update. Can anyone explain why the state parser ignores the existing uuid reference during the import phase? I need the exact syntax to force the provider to accept the remote identifier without regenerating the configuration block.