If I have a job with both partition_by and by_field names defined, does that mean we have models created for all combinations of values for these two fields? How these models are managed (do we keep all of them or gradually remove those that are not seen anymore? ) Is there a way I can track how many models are created for a specific job?
Yes, they are created for every combination and Elastic ML takes care of pruning the models for entities that aren't seen anymore.
You can get a count of how many by_fields and partition_fields exist in the model by looking at the Counts tab for the job at the values of total_by_field_count and total_partition_field_count
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.