Tracking MapReduce job performance with counters
At LiveRamp, most of our heavy data processing is done by MapReduce jobs on our Hadoop YARN cluster. Since these jobs are critical to our data processing workflows, one of our top priorities is making sure they run quickly and reliably.