Forum Discussion

dataengineerv2's avatar
2 months ago
Solved

Fabric Data Factory (REST API): 3 connection blockers migrating a parameterized ADF

 

Migrating an ADF ingestion factory (~219 pipelines, 36 linked services) to Fabric Data Factory programmatically via the REST API. Three connection issues I can't resolve through the API — looking for the supported path or confirmation they're UI/manual-only:
whats is the best apprach to migrate adf pipleins with having parametarized linked services call 
1. Connector types not creatable via the Connections REST API. POST /v1/connections returns "Kind: AzureDatabricks is not supported" and "Kind: Ftp is not supported." Which connector kinds does the public Connections API support, and for the rest (Databricks, FTP, …) is the UI the only option, or is there an automation/API path?

2. ADLS Gen2 + Workspace Identity → DMTS_UntrustedEndpointForWorkspaceIdentity. Creating an AzureDataLakeStorage connection with Workspace Identity fails as an untrusted endpoint. What's required to create it via API end-to-end — does the storage account need trusted-workspace-access / the workspace identity pre-granted? Is there an API to establish that trust?

3. Parameterized linked services with per-call dynamic endpoints. Several ADF REST linked services use url = @{linkedService().absolute_url} — the URL is passed by each dataset at runtime, so one linked service calls many URLs. Fabric connections need a fixed endpoint. What's the recommended pattern — one connection per host + dynamic relative URL? Parameterize the connection field by GUID via a Variable Library? Any way to represent a fully dynamic-URL connection?

Environment: Fabric Trial capacity; connections created via a Service Principal. Official docs/links appreciated.

6 Replies