Ok, the error message clearly states a timeout on the event hub side.
So the best way to approach this is to look at event hub config (instead of going through all kinds of tricks in spark).
Try to set the receiverTimeout/operationTimeout in the connection string, or the maxRatePerPartition.
For the first write, the whole delta table will be read so event hub should be ready for what is in fact a large batch load.
I don't have a clear answer on what setting will work in your situation, so you will have to try out some values (perhaps even combined with spark streaming settings as mentioned before).

event hub config for spark streaming