CubeMX LoRaWAN end node skeleton not retrying OTAA join in event driven mode

As far as I can see, if an OTAA join request does not succeed on first attempt, it will not be retried until an interrupt/event starts another SendTxData callback, which calls LmHandler's LmHandlerIsBusy, which will retry to join, but will return true, which leads to data to be sent being discarded.

On frequent events, that might not be an issue, but for sensors with rare events and bad link quality, this might not be optimal. Should this template be adapted maybe?