{"code": "forbidden_exception", "message": "Access denied. The provided credentials do not have sufficient permissions to perform this action on the specified resource.", "details": "Edge cluster registration token expired or invalid."}
Deployed a new BYOC edge cluster in us-west-2 for a client using AppFoundry. The initial registration via the POST /api/v2/edge/registrations endpoint succeeded, and the cluster status showed as “Active” for roughly 45 minutes. After a scheduled maintenance window required a pod restart, the edge node attempts to re-register automatically. The response is now a persistent 403 Forbidden error.
Verified the OAuth client credentials in the developer portal; they are active and have the correct edge:manage scope. The JWT signature validates correctly against the public key. Interestingly, manual registration attempts from our CI/CD pipeline using the same service account also fail with the same 403, while other API calls (like fetching interaction analytics) work fine.
Is there a hidden state lock on the edge cluster registration after a restart? The documentation mentions a grace period for token refresh, but it does not specify how to reset the registration state if the initial token expires during downtime. We are using the latest Genesys Cloud SDK for Python (v3.2.1).
Has anyone encountered a 403 on edge re-registration after a cluster restart, and what is the correct procedure to reset the registration token without deleting the entire edge cluster configuration?