Looking for advice on token refresh behavior for our multi-org integration.
Have you tried implementing an explicit exponential backoff mechanism within your Node.js token refresh logic? The 502 Bad Gateway after exactly 30 seconds often indicates that the downstream authentication service is timing out due to concurrent load, rather than a fundamental credential error. In my ServiceNow integrations, I avoid hammering the /api/v2/oauth/token endpoint by caching tokens and only requesting a refresh when the expiry window is strictly breached, combined with a jitter-based retry strategy.
Check the payload structure for the refresh_token grant type to ensure it matches the schema exactly. A malformed payload can cause silent failures or gateway errors under load. Refer to the official documentation for the exact parameter requirements here: https://developer.genesys.cloud/api-docs/authentication/oauth2-token-refresh.html. Additionally, verify if your client ID is hitting the rate limit threshold. If the issue persists, inspect the server logs for any upstream timeout indicators from the Genesys Cloud edge nodes, as this often correlates with transient network congestion during peak hours.