When to use Scenes and when to use SceneSections

For the past couple days I’ve been working my way through all the ECS Scene Management samples and the Scenes section of the docs, and I roughly understand how everything works, but I’m still a little confused on when I should use SceneSections.

The 8. Complete example uses them for LODs, which is certainly neat enough and makes sense from an organizational perspective, but ultimately for a level streaming use case where I don’t need LODs it seems roughly superficially equivalent to me that either I bake a bunch of (essentially my own custom meta-) entities with EntitySceneReferences and then runtime load scenes for each chunk as I need them, or I have a single entities scene with just a bunch of section meta entities, that I runtime load the content of for each chunk as needed.

My particular use case is super barebones just like that; I don’t have LODs or cross-chunk-references because everything in the chunks is just AABB colliders and in fact for my built scenario of a server I won’t even have anything to render. I started writing a level streaming system with every chunk in a different entity scene and now I’m wondering if I could or should just use SceneSections.

Looking for insight, further reading, missed gotchas, thanks.

I dont know either. For me it is just overhead and boilerplate to use SceneSections. And SS also brings own restrictions like unloading of SubScene.
Maybe someone can give more information, would also know more about it