The elasticsearch output uses the bulk API to load data. I haven't checked the code, but I would expect it to load each pipeline batch (by default 125 events) as a single API call.
It would not surprise me if rubydebug is your rate-limiting component. How long does it take if you remove the non-elasticsearch outputs.
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.