Metrics aggregation of a sub metrics aggregation

(ben) #1

Hi All,

I require a data visualisation (line chart/vertical bar chart/etc...) that displays something like the following:

X axis should group documents (i.e. top-level buckets) in months over some given time period (lets say "this year" for simplicity)

Y axis should display the average of , the counts of documents of each day , for given top-level bucket (i.e. month)

Is this possible to do in Kibana?

(Mark Walkom) #2

I've moved this to the KB area (it was in ES).

Have you tried anything yet?

(ben) #3

I'm pretty sure this should be in the Elasticsearch area as this is a question about whether this type of search is possible... Kibana only presents what Elasticsearch returns & has nothing to do with the actual analysis / retrieval of data.

I have a meeting set up with someone from Elasticsearch this week to discuss this. The meeting is to discuss 'how to show for each top-level bucket, a metric whose value is dependent on the metrics of the sub-buckets'. I'll be sure to update this topic with any findings.

(Mark Walkom) #4

Yes this is totally possible.
It's just an average with a date histogram on the X axis.

Yeah but you are asking how to do it in Kibana :wink:

(ben) #5

When I choose "average" aggregation for Y-Axis, there is then a drop down box that shows numeric fields. I can only take the average of these numeric fields - there is no option to choose counts of sub-buckets

(system) #6