This service is responsible for executing rules against documents. It receives events on a Kafka topic, determines the appropriate targets (rule and document), executes the rule and records the revisions. Unlike some of our other processing, this service is not a Spark job. We figured there was no reason to make rule execution a Spark job because there is nothing in the execution of a rule that would benefit (directly) from Spark's capabilities while suffering Spark's restrictions on how Scala code should be written (serialization requirements).