Just so I am clear, you are attempting to create 30 individual extracts at the same time from Spark SQL with the Tableau SDK?
Are the 10 tables joined and then adding the 11th causes the .tde creation to fail?
Does attempting the extract with a different 11 or more tables produce any different results or errors?
One suggestion, if you are attempting to create 30 individual extracts,is to just run the job 3 separate times for a each chunk of 10 tables that are needed for the total of 30 tables extracted as this seems to work as expected. Unless there is a need to have them all start at precisely the same time, but that would not guarantee that they would even necessarily finish at the same time, unless I am completely misunderstanding the use case.
Additionally, as I am not familiar with this particular project, I would reach out to the owner of the project by posting to issues section on GitHub.
Feel free to reach out with any clarifying questions.