[Lens] Break down by "all" time series dimensions

[ghudgins]

Describe the feature: For pure mode: time_series data views, allow users to break down by all dimensions--returning every dimension in the data.

Considerations & edge cases:

• Visualization impact: could be a large number of series and risks a lot of DOM elements. Consider elastic-charts issues and smart defaults to mitigate this. The challenge with "other" is many times the insight is derived by seeing the individual series performance and, by grouping 'other', we would miss out and eclipse that in the data. This doesn't necessarily need to be solved in this iteration--there is evidence that users get value out of the "spaghetti" line vis in that the offending series (high value) emerges from the noise. Discuss
• Fallback support: As soon as the context changes to include an index without TSDB then Lens should ignore the breakdown by and return the visualization (not an error) Discuss

Design: Some sort of "All" card in the field list? Something else? Brainstorm / ideas welcome

Describe a specific use case for the feature:
When I have a high value metric that is bad for my business
I need to break down by every single dimension from every field all at once
So I can visually find possible root causes (series driving the high value) without having to check each dimension individually.

[elasticmachine]

Pinging @elastic/kibana-vis-editors @elastic/kibana-vis-editors-external (Team:VisEditors)

[flash1293]

From a technical perspective, the easiest approach would be to have a switch in the "top values" config flyout:

[ghudgins]

we have to consider the bucket limits issue if the user attempts this on high cardinality data

[timductive]

Closing this issue, right now there are similar workarounds so the value here is limited.