We found out, that some deliverer instances were not restarted properly. It took a while to find the issues with a wrong configuration and now as the restart was successful, the sending should perform like normal.
