EDIT: changed in title from "different time-intervals" to "different time-interval lengths"
I would like to display a time-series graph of the percentage of users to whom an event occurred.
This means: the number of users to whom an event occurred during a time interval, divided by the number of "active users" - users who were mentioned in the logs in the 7 days prior to that interval.
The first part: .es(q='event:some_event_name', metrics=cardinality:user.keyword)
The seconds part: .es(metrics=cardinality:user.keyword)
Thank you, but I believe offset still uses the same length of interval as without an offset? As I pointed out, the interval-lengths differ: I'd like to divide a unique count from the last 5 minutes by a unique count from the last 7 days.
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.