[Lens] In-product explanation of how "existing fields" are chosen

[wylieconlon]

We have already made some of the proposed changes from the discussion about the confusing behavior of the field list, but are still getting questions about it that could be better answered in product.

I would propose that we make two changes to Lens:

1. Add a tooltip to the "Available fields" and "Empty fields" text to explain how we calculate this:

We can say "Available fields have data in the first 500 documents in your results"

2. Add a more descriptive text to the field popover, which is covered in a separate issue

[elasticmachine]

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

[wylieconlon]

I think we can explain this using a sentence or two. Here's my suggested text, for "available" and "empty":

Availability is based on the first 500 documents that match your filters. To see all fields, expand the "Empty fields" section. Not all field types can be visualized in Lens, including full text and geographic fields.

These fields did not contain any values in the first 500 documents based on your filters.

cc @KOTungseth @gchaps

[gchaps]

Suggestion:

Available fields have data in the first 500 documents that match your filters. To view all fields, expand Empty fields. Some field types cannot be visualized in Lens, including full text and geographic fields.

[KOTungseth]

I have a few questions/recommendations:

• What if users haven't used the filters? Then shouldn't this stop after 500 documents? If so, I'm still wondering where the 500 documents come from. The first 500 documents of what?
• For the last sentence, could we make it positive? For example, Some field types are unsupported in Lens, including full text and geographic fields.
• For empty fields, I recommend something like Empty fields include data outside of the first 500 documents.

[wylieconlon]

@gchaps I think that suggestion makes sense

@KOTungseth

What if users haven't used the filters?

They don't have to because there is usually a time filter. Not sure we need to say this, but we can.

For empty fields, I recommend something like Empty fields include data outside of the first 500 documents.

The text you're suggesting wouldn't be accurate: all we know is that it's empty, we don't know that there are values.