Developers often make the mistake of-
Hitting
the web service several times by using multiple clusters.
Run
everything on the local node instead of distributing it.
Developers need to be careful with this, as
Spark makes use of memory for processing.