Comment by itaifrenkel

How would you run spark on top of the production db without affecting its performance?