Add histogram aggregate function

[pauldix]

Users should be able to get histograms of values over time. Like this:

select histogram(value, 100) from response_times
group by time(1h) where time > now() - 24h

That query would give you 24 different histograms for each hour where the bucket size of the histogram is 100 milliseconds (assuming that the value column is milliseconds). The response would look something like this:

[{
  "name": "response_times",
  "columns": ["time", "bucket_start", "count"],
  "points": [
    [1384535484, 0, 2],
    [1384535484, 100, 5],
    [1384535484, 200, 23],
    [1384535484, 300, 1],
    [1384531884, 0, 2],
    ...
  ]
}]

So there are multiple points per 1 hour interval where each point represents a bucket in the histogram.

[obfuscurity]

🔥