It looks like we have found the issue now. There was a default read timeout of 60s in our NGINX proxy. After changing that we don't have connection status issues.