histogram count peak - instead of average

[KlavsKlavsen]

I am trying to find peak requests per second - over a larger timeframe (days, weeks, months) - but unfortunately the histogram count - averages the count - so if the resolution in the graph goes up to 30 minutes (which it does for only 2 days search f.ex.) - then I get the total count within the 30 minute windows divided by 1800 (30minutes - in seconds).

What I would like to be able to see - was where I had the highest count-per-second, which would mean that it should take the peak "count-per-second" - instead of averaging.

Is that possible?

[spalger]

This behavior is required in order to keep things performing well and generating comprehensible charts. This is something you could definitely accomplish in elasticsearch, but it's not possible with Kibana. Sorry!