[Lens] Support for all time series functions

[wylieconlon]

This issue organizes the closely related functionality of time series functions.

All time-dependent functions in Lens have the same core behavior, which is that they must be part of a date histogram. Many of these time series functions allow the user to scale to a different unit than the date interval, such as seconds.

Time series functions will most commonly take a metric as input, where a metric is something like "Count of Records". Due to the new implementation plan for calculations in Lens, we can also introduce the possibility of composable inputs such as "Moving average of filter ratio".

There are multiple types of numbers in common use in Elasticsearch:

• Default: Any number without special handling
• Counters: An accumulating number often used in timeseries data, such as network traffic.
• Gauge: Usually represents sensor information such as CPU load or memory usage. Converting these types of numbers into a rate would usually cause confusion.

The following functions are supported: cumulative sum, simple moving average, derivative, counter rate, normalize per time unit

Details

Function	Special logic	Mock
Cumulative sum	Can't allow cumulative sum of cardinality, that's a separate aggregation
Simple moving average	Should support window sizing with previous/next offsets. Allows time scaling.
Derivative	Supports time scaling
Rate for monotonic counter	Supports time scaling
Option to display "per second"	Supports time scaling

The following functions are not supported:

Function	Special logic	Mock
Cumulative cardinality
Linear weighted average
Single-exponential EWMA
Double exponential Holt
Holt-Winters
Percent growth
Instant delta	Should used for gauge. Compares the last 2 values in the time range

Technical requirements:

• Per second suffix formatter
• Expression function to do time scaling

[elasticmachine]

Pinging @elastic/kibana-app (Team:KibanaApp)

[flash1293]

Closing this as there are issues for all unresolved parts of this already - tracked in #57708