Minor Description
Currently, by default, every second a job is run a massive JSON string containing the query plan, the tasks, and some counters is logged to the hive_job_log. For large, long running jobs that can easily reach gigabytes of data. This logging should be configurable as average user doesn't need this logging.