What I want to get out of the data is what is the 60 minute period of time with the highest sum(Concurrency). Looking simply, I could do a | stats sum(Concurrency) as HourlySum by date_hour | stats max(HourlySum). The only problem there is that it cuts the window directly along 00:00-59:59 minute periods. If the busiest period of time is 13:22:00-14:21:59, it wouldn't be properly represented.
Is there any way to do a rolling average, so I could get the data I need out of it?