[Lens] Selected date range should contain all computed moving average data points

[dej611]

Describe the feature:

When configuring a moving average for a given date range the first and last buckets are not presented as they are used to compute the initial and final moving average data points:

A customer asked if it would be possible to have the selected range full of data points for the moving average operation.
This would require Lens to operate in a different way based on the operation type, which would not be ideal, but maybe there's another way.

[elasticmachine]

Pinging @elastic/kibana-visualizations @elastic/kibana-visualizations-external (Team:Visualizations)

[markov00]

Adding a buffer before and after the time could be definitely interesting. The chart could still "clip" the time to the specified time range, but the chart data could extend outside the current range and to keep a correct representation of the trend even behind the ranges.
This goes in the direction of providing panning capabilities like the timeslip demo where we want to load a bit more data before and after the current time range to allow a smoother interaction when panning.

[dej611]

The only problem I see with this is the bucketing estimation for the date histogram: while there's now a way to have a pretty accurate estimation, increasing the time range would potentially change the bucket size by ES, producing a mismatch between the presentation and the actual table. Lens could enforce a bucket size in this case when in Auto, but it doesn't sound like a solid solution to me.

[markov00]

In order to provide better transparency of priorities, issues that will not be prioritized within the next 24 months are being closed.

Tracking request in Lens general improvements ice box #184459