Query Builder: Add ability to group by columns

[hoorayimhelping]

Related: #15563

It is currently only possible to generate queries of the form

from(bucket: "telegraf")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(<filter>)
  |> aggregateWindow(every: v.windowPeriod, fn: <agg>)

Since flux does a default group by *, it is not possible to apply a different grouping of data before aggregating.

1. Aggregate all series over time into a single series

More of less the query we'll want to be able to express is

from(bucket: "telegraf")
  |> range(start: v.timeRangeStart, stop: v.timeRangeStop)
  |> filter(fn: (r) => r._measurement == "cpu")
  |> filter(fn: (r) => r._field == "usage_user")
  |> group()
  |> aggregateWindow(every: v.windowPeriod, fn: mean)
  |> yield(name: "mean")

[hoorayimhelping]

POC I whipped up:

Chatted with Alex Pax about this on Slack. We both think there needs to be more affordance for switching between Filters and Groups so I'm going to make the Filter / Group a dropdown and allow selecting it.

[hoorayimhelping]

Chatted with @mavarius about the design and UX of this. For the first initial pass (what I'm calling Group By 0.5) we're going to add a dropdown to select between filter and group. Adding a new BuilderCard will select filter by default, initially.

After the initial 0.5 functionality lands, we'll evaluate the best way to change the function with fewer clicks. Some initial thoughts are to have a list of functions to add in place of the single + icon, or a dropdown above Aggregate Functions with a list of functions to add.

[hoorayimhelping]

Related: #16076