Hi,
Apologies about the wall of text, but there are a lot of good questions in your thread I want to help with.
First, you are correct that a private Genesys Cloud Embeddable Framework deployment is subject to API usage allocations.
The important distinction here is that API usage is measured on the Platform API requests made by the OAuth client, rather than on the individual GCEF method, action, or event itself. The API Usage view records the PUT, POST, GET, and DELETE requests made by OAuth clients.
Because of that, there isn't a documented fixed request count that I can give you for initialSetup, clickToDial, Interaction.addCustomAttributes, addTransferContext, User.updateStatus, Interaction.updateState, or screenPop. Genesys doesn't document a fixed number of Platform API requests for each GCEF action, so there isn't a reliable per-action request count we can provide.
The same distinction is important for subscriptions. A callback from an Interaction or UserAction subscription isn't itself equivalent to another REST API request through the OAuth. I wouldn't model API allocation as "one request per event received." What matters for the allocation is the actual Platform API traffic generated by the client.
For the built-in GCEF UI, I also wouldn't assume that an agent action such as Answer, Hold, or Transfer is exempt simply because it originated from the embedded UI. Genesys specifically calls out Embeddable Framework applications as being subject to API usage allocations, unlike the Genesys Cloud browser/web/mobile apps and the explicitly excluded pre-built integrations.
Configuration can certainly affect overall API traffic because it can change the behavior exercised by the framework, but there isn't a supported formula such as "each custom attribute adds one request" or "enabling this setting adds X requests." I wouldn't recommend building a capacity model around assumptions at that level.
For your pre-production (sandbox) testing, I recommend using a dedicated OAuth client in your test environment, exercise the same GCEF configuration and workflows you expect in production, and use the API Usage view to measure the requests generated by that client. The view lets you drill into both the OAuth client and the individual API requests, which should give you a much more reliable baseline than trying to derive request counts from the GCEF JavaScript methods. If you need more info on the API usage view, please see the documentation below:
API Usage view
About APIs and API usage
There is a lot to unpack here, so here is the summary of the above:
Genesys Cloud emebbedable framework is subject to API usage allocations through the OAuth Requests, but there isn't a documented per-GCEF-operation request-count contract. For sizing, measure the actual Platform API traffic generated by your private deployment and configuration rather than assuming a fixed mapping between GCEF actions/events and API requests.
Hope this helps!
------------------------------
Cameron
Online Community Manager/Moderator
------------------------------