Title: Support dbt-fabricspark adapter (Microsoft Fabric Lakehouse via Livy)
Is your feature request related to a problem? Please describe.
Rosetta DBT Studio currently supports a fixed set of connection types (postgres, snowflake, bigquery, redshift, databricks, mysql, oracle, db2, mssql, kinetica, googlecloud, duckdb, ducklake) and a matching allowlist of installable dbt adapter packages (dbt-postgres, dbt-snowflake, dbt-bigquery, dbt-redshift, dbt-databricks, dbt-duckdb). There's no path to using Microsoft Fabric Lakehouse (Spark) as a target — it's not in SupportedConnectionTypes, and dbt-fabricspark isn't in the package allowlist, so it can't be installed through the app's package manager even manually.
Teams building on Microsoft Fabric who want to use dbt Core for their Lakehouse transformations currently have no way to do that from within Rosetta DBT Studio and have to fall back to plain dbt Core via CLI.
Describe the solution you'd like
Add dbt-fabricspark (https://github.com/microsoft/dbt-fabricspark) as a supported adapter/connection type, including:
- A
fabricspark entry in SupportedConnectionTypes and a corresponding connection form (workspace name, lakehouse, schema, authentication method — Azure CLI / service principal / etc., matching the adapter's profiles.yml options)
dbt-fabricspark added to the adapter package allowlist so it can be installed/managed like the existing adapters- Profile generation support (equivalent to the existing per-adapter cases in the connectors service) that emits a valid
fabricspark profile block
Describe alternatives you've considered
- Using
dbt-databricks as a substitute — doesn't work, since Fabric Lakehouse is accessed via Livy endpoints rather than the Databricks API/connection model, so the connection semantics don't match.
- Manually installing
dbt-fabricspark into the app's bundled Python environment outside the UI — even if technically possible, the app's connection setup, model generation, and AI-assisted features are all built around the fixed connection-type list, so this would bypass most of the tool's value rather than integrate with it.
Additional context
dbt-fabricspark is Microsoft-maintained, connects via Livy, and supports schema-enabled and non-schema Lakehouse configurations along with table/view/incremental/seed/snapshot materializations — so it's an actively maintained adapter, not an edge case. Happy to help test against a real Fabric workspace if that's useful during development.
Title: Support
dbt-fabricsparkadapter (Microsoft Fabric Lakehouse via Livy)Is your feature request related to a problem? Please describe.
Rosetta DBT Studio currently supports a fixed set of connection types (
postgres,snowflake,bigquery,redshift,databricks,mysql,oracle,db2,mssql,kinetica,googlecloud,duckdb,ducklake) and a matching allowlist of installable dbt adapter packages (dbt-postgres,dbt-snowflake,dbt-bigquery,dbt-redshift,dbt-databricks,dbt-duckdb). There's no path to using Microsoft Fabric Lakehouse (Spark) as a target — it's not inSupportedConnectionTypes, anddbt-fabricsparkisn't in the package allowlist, so it can't be installed through the app's package manager even manually.Teams building on Microsoft Fabric who want to use dbt Core for their Lakehouse transformations currently have no way to do that from within Rosetta DBT Studio and have to fall back to plain dbt Core via CLI.
Describe the solution you'd like
Add
dbt-fabricspark(https://github.com/microsoft/dbt-fabricspark) as a supported adapter/connection type, including:fabricsparkentry inSupportedConnectionTypesand a corresponding connection form (workspace name, lakehouse, schema, authentication method — Azure CLI / service principal / etc., matching the adapter'sprofiles.ymloptions)dbt-fabricsparkadded to the adapter package allowlist so it can be installed/managed like the existing adaptersfabricsparkprofile blockDescribe alternatives you've considered
dbt-databricksas a substitute — doesn't work, since Fabric Lakehouse is accessed via Livy endpoints rather than the Databricks API/connection model, so the connection semantics don't match.dbt-fabricsparkinto the app's bundled Python environment outside the UI — even if technically possible, the app's connection setup, model generation, and AI-assisted features are all built around the fixed connection-type list, so this would bypass most of the tool's value rather than integrate with it.Additional context
dbt-fabricsparkis Microsoft-maintained, connects via Livy, and supports schema-enabled and non-schema Lakehouse configurations along with table/view/incremental/seed/snapshot materializations — so it's an actively maintained adapter, not an edge case. Happy to help test against a real Fabric workspace if that's useful during development.