@RyanDeBruynThe documentation says that it's sorting by the "Best" field, which relates to smallest pixel size.
However, that refers to the mosaic and "display" of the data. Does that sorting also apply when the service is used for analysis, such as a geoprocessing tool?
What I'm trying to get at is it seems like there's no way to be confident about the values the Terrain service returns without using a definition query. For example, if the spatial extent of my vector data is larger (even by a single small feature!) than the extent of the underlying "Best" part of the service, then it is unclear what the results are. Does the tool apply "Best" to the features it covers, and then "Second Best" to a feature beyond its extent? Or is the entire area reduced to "Second Best" to ensure coverage?
Sometimes when I used the FMV Video Multiplexor, the tool fails due to raster input errors with the ESRI Terrain service. I think the specific tool here isn't so important. Just trying to understand what's actually happening with the data.