In SumoLogic we created a visualization that took an average of all the execution times and filtered out the top 10%, so you could see a truer average of processtime rather than a couple of outliers creating spikes. The where part of our query looked like -
| pct(executiontime,90) by _timeslice, processtype
I can't figure out how to get the same results in Kibana. Any thoughts?
Hey @Alan_Home, this isn't possible currently using the "standard" Visualizations in Kibana. Implementing this would require two calls to Elasticsearch, one to calculate the 90th percentile and then another query to calculate the average, but excluding all values that are above this limit.
It's theoretically possible for you to create this yourself using Vega but it's a rather heavy lift. It'd be great to also capture this requirement in a feature request so we can address it more holistically.
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.