SQL Server 2008 on-prem - CDC Fabric

Dears,

My architecture in fabric is :

Raw->Bronze (with history)->Silver->Gold

The topic of this post is related with Raw and Bronze

In Raw, we are consuming different data sources.

For example, we have a Dataverse Direct Link adding information into our Raw.

1) Behavior of Direct Link:

There is a Dataverse Direct Link created at dataverse side which continuously sends shortcuts into our RAW.

The DV Direct link creates, by itself, a lakehouse in our raw workspace and keeps it up to date with all the tables 

Information keeps arriving and when there is something deleted, it has a flag called isdelete (so that we are aware of deletes at source)

Then, every 30 days, those deletes get a hard delete from this lakehouse in our raw. All of this managed by Dataverse Direct link.

Then, In bronze, we do incremental loads from this raw tables using a column which is called SynkmodifiedOn, hence we always have all the history in bronze including the deletes

The challange is how to consume deletes in the same way from SQL Server 2008 DB 

2) Now we have new data source, its SQL Server 2008 on-prem

We are considering doing CDC, and we would like to have the same capacity in terms of knowing when something was deleted

I was thinking in doing something like this:

In SQL Server :

1) Activate CDC at DB level and table level in SQL Server

2) Ensure SQL Agent is running

3)Configure CDC retention and Monitoring capture/cleanupjobs for 7 days period

In Fabric, 

Create a connection to SQL Server (via a on-premisses data gateway instealled on a VM that can access the SQL Server)

Create a copy job and select CDC mode

Create a lakehouse in Raw 

Point the copy job to it

Schedule the Copy job to run 1 a day (we don't need more at this time)

Authentication (seems that because the SQL Server is on-prem) I cannot use a service principal , I have to use windows authentication via the gateway or SQL authentication

But I heard that this way the copy job will not be able to bring into raw the information of the records that were deleted.

This is challeging because we would like to have something like a flag that we can use to know when something was deleted

Can someone please help?  Maybe I should not use the copy job to bring the infromation ? should instead use a copy task from Fabric Data Factory?

Do you think the above list of requirements is correct? or am i lacking something?

Thanks a lot for an note

Best regards,

Pedro