Good morning,
in our project we are trying to optimize the performance of our point vector layers represented with SVG styles, migrating the logic to WebGLPointsLayer type layers.

In the presented use case we see different numerical values associated to the asset.

We know that to migrate correctly to WebGLPointsLayer we should previously use a sprite to assign the style to each element. Therefore, we should know the content of the Source and attributes of the different features that will generate each element of the sprite. In layers with a large number of elements, excessively large sprites can be generated. We can talk about layers of up to 100,000 points and several hundred different styles.

We think that the process we should follow is:
1-. Obtain features associated with the source and differentiate components of the sprite.
2-. Generate the sprite from the components analysed.
3-. Create the WebGLPointsLayer from the source obtained in point 1-. and the styles generated in point 2-.

@jahow , I have seen that you are quite experienced in this area.

Is this the correct way in which you would face this situation?
Do you know any alternative to optimize it?
Is there, or is it expected, the possibility of including texts on the icons?

Best regards and thank you very much