I’m loving the new Histogram panel, and it works beautifully with raw data from InfluxDB. I’m starting to look into the Prometheus Histogram datatype, and it seems this could be a good fit for the “pre-bucketed data” option in the Histogram panel.
I’m not sure how I can structure my Prometheus query to satisfy this requirement though:
> Each time series already represents a Y-Axis bucket. The time series name (alias) needs to be a numeric value representing the upper interval for the bucket. Grafana does no bucketing so the bucket size options are hidden.
So interestingly enough i was actually using those, trying to build a basic panel to help an internal user, and despite my best efforts I always came out with something where the Heatmap showed the le buckets out of order. And for the histogram panel would show count or values rather than the actual histograms.