I am trying to keep my credit usage minimal on ArcGIS Online. I have a standard python notebook that reaches out to an API, makes a query , pushes the query results into a spatial dataframe and then either updates a feature layer if it already exists, or if not, publishes the feature layer. Everything works, its great.
However, I noticed my credit balance dropping a little faster than I'd like. So I wanted to take a look at whether I was using the most appropriate feature layer type. I took a look at the Understanding Credits documentation, specifically Credits by Capability and noted that Feature Storage is charged at a rate of 2.4 credits per 10MB stored, calculated hourly while Feature Collection storage is charged at a rate of 1.2 credits per 1GB stored, calculated hourly.
That's a significant savings, but I am confused.
Feature Storage cites the example of storing a hosted feature layer - makes sense. Meanwhile, storage of all content, excluding hosted feature layers and content in ArcGIS Notebooks cites multiple different examples, including storing Feature Collections.
In my python notebook, I create a Feature Layer from a spatial dataframe already populated with x/y coordinates and attributes using:

which produces a Feature Layer Collection. All is good. But then I look in My Content in my organization and the item type is shown as Feature Layer (hosted).

So which is it? A Hosted Feature Layer or a Feature Layer Collection? Am I going to be charged 2.4 credits per 10MB/month or 1.2 credits per 1GB/month for my Feature Layer Collection/Feature Layer (hosted) item.
OR am I thinking about Feature Collections wrong and the Feature Layer collection is actually the Shapefile that was generated, which the Feature Layer (hosted) was then created from and the Shapefile is what costs 1.2 credits per 1GB/month while the Feature Layer (hosted) costs 2/4 credits per 10MB/month.