Got the connection open, auth header is valid, and the initial subscribed event comes through fine. But the actual conversation:updated payloads just stop after about 30 seconds of inactivity on the org side. The socket doesn’t close, no error event fires, it just goes silent.
ws.send(JSON.stringify({
"type": "subscribe",
"events": ["conversation:updated"],
"filter": { "orgId": "my-org-id" }
}));
I’m running this in a Node.js worker. The server is Genesys Cloud. I’ve checked the retry logic on my end, but the issue is the source stream drying up. Is there a specific ping/pong requirement I’m missing, or does the Notification API have a hard idle timeout that silently drops the sub? The docs mention a heartbeat but don’t specify if the client needs to initiate it. Tried adding a manual ping every 15 seconds, didn’t help. Any ideas?