Packages and deployments
Review .dtsx, .dtproj and .ispac assets, deployment models, parameters, environments and shared resources.
Enterprise ETL modernization
SSIS to Databricks Accelerator: Modernize Enterprise ETL Workflows
Plan the transition of SSIS packages, data flows, scripts, schedules and dependencies into Databricks-native data engineering patterns.
Start with the full operational estate, choose fit-for-purpose Spark and orchestration patterns, then rebuild and validate workflows before controlled cutover. SSIS packages do not run natively on Databricks, and the right migration path depends on the package logic and its runtime dependencies.
Accelerator coverage and migration outcomes depend on the source estate and agreed scope; confirm product capabilities with the team.
Estate discovery
What should an SSIS migration assessment cover?
A .dtsx package is only one part of a production workflow. Include deployment, scheduling, external dependencies and operational controls before designing the Databricks target.
Control and data flow
Document tasks, containers, precedence constraints, transformations, paths, error outputs and restart behavior.
Runtime dependencies
Identify connection providers, drivers, custom components, scripts, stored procedures, files and external services.
Production operations
Inventory SQL Server Agent jobs, schedules, proxies, retries, notifications, SLAs and downstream dependencies.
Design patterns
Map SSIS components to Databricks deliberately
These are possible target patterns, not guaranteed one-to-one conversions. Confirm connector support, execution behavior, security and business semantics for each workload.
| SSIS component or concept | Possible Databricks target | Migration consideration |
|---|---|---|
| Control Flow and precedence constraints | Lakeflow Jobs task graph or another orchestrator | Rebuild branches, conditions, dependencies and failure paths explicitly. |
| Data Flow Task | PySpark, Spark SQL or a suitable pipeline | Reimplement source-to-destination behavior and validate edge cases. |
| Lookup, Merge Join and Union All | Spark joins and unionByName | Check duplicate handling, nulls, key behavior and schema coercion. |
| Derived Column and Data Conversion | Spark SQL expressions or PySpark columns | Test casts, precision, truncation, date functions and invalid values. |
| Aggregate and Sort | Spark aggregations, orderBy or window logic | Define grouping grain and only rely on ordering where required. |
| Slowly Changing Dimension | Explicit Delta MERGE and dimension logic | Specify Type 1/Type 2 rules, effective dates and late-arriving data. |
| Script Task and Script Component | Python package, notebook, PySpark or retained service | Analyze .NET dependencies, side effects, metadata and error behavior. |
| Execute SQL Task and stored procedures | Spark SQL, PySpark, retained database procedure or service | Review dialect, transactions, temporary tables and engine-specific logic. |
| Connection managers and parameters | Supported connectors, job parameters and approved secret management | Rebuild identity, network access, configuration and environment boundaries. |
| SQL Server Agent schedules and event handlers | Lakeflow Jobs or an enterprise scheduler with logging and alerts | Preserve calendars, retries, notifications and operational ownership. |
Delivery approach
SSIS-to-Databricks migration methodology
- Phase 1
Discover the SSIS estate
Inventory packages, SSISDB or legacy deployments, SQL Agent jobs, dependencies, custom code, data sources and owners.
- Phase 2
Assess complexity and readiness
Classify transformations, control flow, provider dependencies, data volumes, criticality and candidates for retirement.
- Phase 3
Design the Databricks target
Choose ingestion, Delta storage, Spark transformations, orchestration, governance, security and acceptance criteria.
- Phase 4
Prove representative patterns
Pilot simple and complex packages, branching, custom code and schedule-dependent workflows; record exceptions and manual effort.
- Phase 5
Rebuild and validate
Implement workflows and data models, then test source-to-target results, error handling, recovery and operational behavior.
- Phase 6
Cut over in controlled waves
Use dependency-aware releases, business acceptance, rollback gates, runbooks and support handover before retiring SSIS.
Acceptance criteria
Validate behavior before cutover
A successful code rebuild does not alone prove that data meaning, failure handling or production operations were preserved. Agree acceptance checks and aligned source snapshots with the business owners.
- Schemas, data types, row counts, business keys and aggregate comparisons
- Nulls, duplicate behavior, ordering assumptions and error-row handling
- Incremental loads, retries, idempotency and restart boundaries
- Control-flow branches, failure paths, schedules and notifications
- Credentials, network access, permissions and audit requirements
- Representative runtime, concurrency, storage and total-cost measurements
Common questions
SSIS to Databricks migration FAQs
Can SSIS packages run directly on Databricks?
Databricks is not an SSIS runtime. SSIS logic generally needs to be rebuilt using target-native transformations and orchestration. Azure-SSIS Integration Runtime is a separate option for running compatible packages in Azure Data Factory.
Is SSIS-to-Databricks a lift-and-shift?
Usually not. Package behavior, dependencies, data semantics and operations need to be assessed and rebuilt for the target execution model.
Can .dtsx files be converted automatically?
Automation may help with package parsing, dependency discovery, classification and supported patterns. Complex control flow, custom code and provider-specific behavior still require engineering review and testing.
What happens to SSIS Control Flow and Data Flow?
Control Flow can be reconstructed as a task graph or another orchestrator. Data Flow transformations can be redesigned with PySpark, Spark SQL or suitable pipeline components. Conditions, error paths and data semantics must be tested.
Can SSIS Script Tasks be converted to Python?
Some logic can be rewritten, but arbitrary .NET code has no guaranteed direct conversion. Review libraries, runtime dependencies, file and network access, side effects and error behavior.
What happens to SQL Server Agent jobs and SSISDB environments?
Inventory schedules, dependencies, retries, proxies, parameters, environment references and permissions. Recreate the required configuration and operational controls in the target rather than copying sensitive values.
How are migrated packages validated?
Compare aligned source and target data, including schemas, counts, keys, aggregates, nulls, incremental changes and errors. Also validate schedules, failure handling, runtime and business outcomes against agreed criteria.
Does this page guarantee automatic conversion or a specific effort reduction?
No. The precise scope of any SSIS-to-Databricks accelerator must be confirmed with the product team. Package coverage, automation, effort, cost and timing depend on the source estate and should be established through assessment and a representative pilot.
Office Solution AI Labs
Plan your SSIS-to-Databricks modernization
Scope a migration around package inventory, dependency analysis, target architecture, workflow rebuilding, data reconciliation and cutover readiness. Confirm supported accelerator patterns and acceptance evidence for your estate before setting expectations.
