I'm planning to use Composite Aggregation in one of my projects and wanted to know how expensive these operations are. Specifically, I wanted to see how does the compute and memory cost increase along with following values.
Total cardinality of the query
Total number of documents in the cluster/shard/index
Addition of filters
Is there any downside in using the Composite Aggregation?
Most uses of the composite aggregation only visit the docs being aggregated in the current pass. The docs that don't "fit" in the range returned by the composite are mostly skipped. So generally the cost of any individual execution of composite is proportional to the number of docs with terms that fall in the next n composite sort keys and not propositional to the total number of docs in the index or that match the query.
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.