I have some data I am querying from Prometheus, a range vector, that I am currently getting via sum by (label1, label2) (series_name), and as expected, it gives me the most recent results of the sums by the labels. I am trying to change it to get the results over a timeframe (let’s say 1 minute with a resolution of 5 seconds).

I’ve tried to use sum_over_time by (label1, label2) (series_name[1m]) but aggregating with by doesn’t seem supported by aggregation over time, so it just says that by was unexpected.