Multiple timelines?

Hopefully my noobness isn’t too offensive.

So I’ve managed to successfully load my data to Elasticsearch
And have been meddling with Grafana, i can show graphs …proving my data is present.

The problem I’m having is my data is like this…
start: 12312513423( epoch start )
end 123128376823( epoch end )
bandwidth 123 (number )

So ideally i want to graph a 5 minutely summary of the bandwidth used. So it would total up all bandwidth that falls between start and end to the time point selected.

I’ve tried to google for some assistance but not entirely sure of the term to search for.
Do i need to have two "time field name"s?

Or, which i suspect is the answer, do i need to add something to the Query ( currently doing Graph, query metric of sim 'bandwidth). group by date histogram.

If someone could point me in the right direction or throw the syntax sample here I’d much appreciate it.

Thanks for your time!


Normally you would have one timefield (date/time of event) and the you aggregate (group by) date histogram and set what interval you’re interested in - that allow you to see the bandwidth used per interval bucket, like bandwidth per 5 minute or 1 minute or 1 hour etc.


Yeah, thats where I’m struggling to get my head around this.
My actual data source gives me Start time (unix time) and a duration (seconds) and the total KB’s written in that time. So rather un-usefully potentially vast amounts of kb all associated with one specific time. I thought it useful to translate that into 2 additional fields of end time and bandwidth to make this a bit more doable.

Now i kinda get the pseudo logic of select sum(bandwidth) from $table where $time > start and $time < end
but just can’t figure this out. Maybe this just isn’t meant to be. Thanks for your time mefraimsson