Mask timeslider

[nreese]

Client side masking was removed from original timeslider implementation.

This PR brings back client side masking to timeslider with the following improvements

• masking field is pulled from source data request meta so ILayer.syncLayerWithMB does not need to make async
• ESSearchSource.getGeoJsonWithMeta performs a count request prior to fetching documents to know if it can pull for the entire time range or just the timeslice.

[elasticmachine]

Pinging @elastic/kibana-gis (Team:Geo)

[thomasneirynck]

thx!

most of this was discussed offline with @nreese, but for completeness.

The current approach is hard to generalize to .mvt based data. For mvt-data, the source cannot determine up front whether all data is on-screen. Instead, it needs to inspect the actual responses of the tiles (reading metadata from tiles is being worked on here #94811).

As next steps, we will introduce in this PR:

• pull up determining whether data is complete into a layer-method. This can be implemented differently, based per layer type (in practice, there will be a different implementations for vector_layer and tiled_vector_layer).
• source could stilll make a pre-fetch call to fetch all data for the entire-range, rather than just the timeslice. This would mean that es-sources with the LIMIT-scaling type can still prefetch all data for the entire range (rather than just the timeslice). This is an optimization that would only work the LIMIT-scaling-type. We determine it is valuable enough to release, mainly because a lot of the mvt-work on the ES/and Kibana sides is still in flux ([Maps] [Meta] .mvt feature completeness #79868, Support .mvt vector tile format elasticsearch#58696).

[nreese]

@elasticmachine merge upstream

[spalger]

jenkins, test this

(restarting due to jenkins upgrade)

[nreese]

@elasticmachine merge upstream

[thomasneirynck]

Thx, +1 on overall changes. I think we should be able to hack on tile-based data here as well (tbd).

One small bug I believe, see comments.

[thomasneirynck]

thanks, this is a great improvement. It will make timeslider work for even more datasets.
See one comment to improve readability.

[nreese]

@elasticmachine merge upstream

[kibanamachine]

💚 Build Succeeded

• continuous-integration/kibana-ci/pull-request
• Commit: 65494d7
• Storybooks Preview

Metrics [docs]

Async chunks

Total size of all lazy-loaded chunks that will be downloaded as the user navigates the app

id	before	after	diff
maps	3.1MB	3.1MB	+4.1KB

History

• 💚 Build #133519 succeeded 24cbb55
• 💚 Build #132699 succeeded 6c1a614
• 💔 Build #132675 failed 78b38f5
• 💚 Build #131880 succeeded 58253e2
• 💚 Build #131867 succeeded d8de9e4

To update your PR or re-run it, just comment with:
@elasticmachine merge upstream

[kibanamachine]

💔 Backport failed

Status	Branch	Result
❌	7.x	Commit could not be cherrypicked due to conflicts

To backport manually run:
node scripts/backport --pr 102046

[nreese]

7.x backport #103132