It would be good to automatically rerun a transformation job that failed to run. 

Recently we ran into a problem where we had set two different transformation jobs to start at the same time. There were cases where this caused problems because both of them were issuing a "dbt seed". In rare cases, the first job was reconstructing the table (seed) and at the same time, the second job was looking for this table.

The table was missing at the time and the job was failing. 

If there was a setting to select to rerun the failed job automatically, that would be great addition.