I am trying to optimize my hardware for elasticsearch deployment. I am getting inconsistent results on the time by changing the JVM settings. I wanted to know if there is any other measure apart from the "took" in the result query to capture the time the query spends on the hardware.
Also, is there a way to capture the time spent by the query in Disk I/O.
Finally, is there a method of getting detailed time breakdown for the query?
You can use the Profile API to get a detailed breakdown of query performance. I'm not sure if there is a way to see Disk I/O directly related to a query as this becomes a more complex topic when you take into things like filesystem level caching and other things. The only way to really do this would be to use something like rally to run at scale benchmarks and make changes to hardware to see what impacts they have.
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.