Anytime you want to measure distributed processes, you should look into 
logging, or sending the time data asynchronously to a persistent data store. I 
haven’t measured joinWithCassandra but I’ve measured other parts of spark using 
a Kafka topic to send execution times. I then consumed the times and saved them 
into Cassandra. I could then later get time aggregates and average times per 

Rahul Singh

Anant Corporation

On May 13, 2018, 4:14 PM -0500, Guillermo Ortiz <>, wrote:
> I'm using the driver from Cassandra-Spark, I would like to know if there is a 
> way to measure the time which takes a " joinWithCassandraTable" to execute 
> easily inside a spark code.

Reply via email to