Data Factory Salesforce connector changing. What should you do?

In the last few days, Microsoft released a new version of the Azure Data Factory Salesforce connector. The old connector is deprecated, and you can no longer create a new Linked service or datasets of it, but you can continue to use existing objects of the old connector, with a nice "this connector is legacy" message on them.

What are the differences?

	Legacy connector	New connector
Salesforce API	REST/Bulk API	Bulk API 2.0
Authentication	User and password	OAuth 2.0
Documentation	https://learn.microsoft.com/en-us/azure/data-factory/connector-salesforce?tabs=data-factory	https://learn.microsoft.com/en-us/azure/data-factory/connector-salesforce-legacy?tabs=data-factory

Here is a comparison of the "create linked service" window:

What should you do if you are using or want to use this connector?

New environments:

You cannot create Linked services for the old salesforce connector, so you will have to use the new one.

There is a workaround if you are moving objects between environments, for example, from dev to prod, you can use an ARM template to copy objects between them, and include the old connector.

Existing environments:

You can continue to use existing Linked services and datasets of the legacy connector and make changes to them. You cannot create a new Linked service for the legacy connector, and Microsoft advise to create future objects with the new connector.

If you try to create a new dataset for the salesforce connector, you will only see the new version Linked service (if exists) and cannot connect the dataset to the legacy connector.

There is a workaround though. You can clone existing datasets and then change them as required to get the new dataset you need.

For the moment, the legacy connector is supported, and Microsoft did not announced an end of support date, but since they do not let customers create new instances of this connector, we can assume they will remove it in some point in the future. If you can, move to the new connector. If you have a lot of datasets, this may be a project, because you need to replace each dataset wherever it is used, but you will gain improve native Salesforce support (according to Microsoft), and peach of mind that your connector will not work in the future is important too.