Azure Data Factory Databricks Integration 2025
Databricks Job Activity (Recommended 2025)
CRITICAL UPDATE (2025): The Databricks Job activity is now the ONLY recommended method for orchestrating Databricks in ADF. Microsoft strongly recommends migrating from legacy Notebook, Python, and JAR activities.
Quick Reference
- Activity type:
DatabricksJob (NOT DatabricksSparkJob or DatabricksNotebook)
- Parameter property:
jobParameters (NOT parameters)
- Linked service auth: Managed Identity (
"authentication": "MSI") recommended
- Cluster config: Do NOT specify cluster properties in linked service; the Databricks Job controls compute
Why Databricks Job Activity?
| Feature |
Notebook Activity (Legacy) |
Job Activity (2025) |
| Compute |
Must configure cluster in linked service |
Serverless by default |
| Workflow tasks |
Single notebook |
Multi-task DAGs (notebook, Python, SQL, DLT) |
| Retry |
ADF-level only |
Job-level + task-level |
| Repair runs |
Not supported |
Rerun failed tasks only |
| Git integration |
Limited |
Full Databricks Git support + DABs |
| Lineage |
None |
Built-in data lineage |
| If/Else logic |
Must use ADF control flow |
Native If/Else task types |
Benefits Summary
- Serverless Execution -- No cluster specification needed; automatic serverless compute with faster startup and lower costs
- Advanced Workflow Features -- Run As, Task Values, Conditional Execution, AI/BI Tasks, Repair Runs, Notifications, Queuing
- Centralized Job Management -- Jobs defined once in Databricks workspace; single source of truth with Git-backed versioning
- Cost Optimization -- Serverless compute (pay only for execution), job clusters (auto-terminating), spot instance support
For complete JSON examples of Job activity, linked service, and pipeline configurations, see references/databricks-job-examples.md.
Connectors and Enhancements (2025+)
ServiceNow V2 Connector (RECOMMENDED - V1 End of Support)
ServiceNow V1 connector is at End of Support. Migrate to V2 immediately.
| Feature |
V1 |
V2 |
| Linked service type |
ServiceNow |
ServiceNowV2 |
| Source type |
ServiceNowSource |
ServiceNowV2Source |
| Query builder |
Custom |
Aligns with ServiceNow condition builder |
| Performance |
Standard |
Enhanced extraction |
| OData support |
No |
Yes |
Migration steps: Update linked service type to ServiceNowV2, update source type to ServiceNowV2Source, test queries in ServiceNow UI condition builder, adjust timeouts.
Enhanced PostgreSQL Connector
Improved performance with 2025 SSL enhancements: enableSsl: true, sslMode: "Require".
Enhanced Snowflake Connector
Improved performance with KeyPair authentication support and Key Vault secret integration.
Managed Identity for Azure Storage
New managed identity support for Azure Table Storage and Azure Files connectors (system-assigned and user-assigned).
Mapping Data Flows - Spark 3.3
Spark 3.3 now powers Mapping Data Flows with 30% faster processing, Adaptive Query Execution (AQE), dynamic partition pruning, improved caching, and better column statistics.
Azure DevOps Server 2022 Support
Git integration now supports on-premises Azure DevOps Server 2022 via the hostName property.
For complete JSON examples of all connectors, see references/connector-examples.md.
Managed Identity 2025 Best Practices
User-Assigned vs System-Assigned
| Scenario |
Recommendation |
| Single ADF, simple setup |
System-assigned |
| Multiple data factories |
User-assigned (shared identity) |
| Complex multi-environment |
User-assigned |
| Granular permission control |
User-assigned |
| Identity lifecycle independence |
User-assigned |
Use ADF's centralized Credentials feature to consolidate Microsoft Entra ID-based credentials across multiple linked services.
MFA Enforcement (Enforced Since October 2025)
Azure MFA is mandatory for all interactive user logins. Impact on ADF:
- Managed identities are UNAFFECTED -- no MFA required for service accounts
- Service principals with certificate auth are the recommended alternative to secrets
- All interactive user logins require MFA
Principle of Least Privilege
| Resource |
Source Role |
Sink Role |
| Storage Blob |
Storage Blob Data Reader |
Storage Blob Data Contributor |
| SQL Database |
db_datareader |
db_datareader + db_datawriter |
| Key Vault |
Get secrets only |
Get secrets only |
For complete managed identity JSON examples, see references/connector-examples.md.
Best Practices (2025)
Use Databricks Job Activity (MANDATORY) -- Stop using Notebook, Python, JAR activities. Define workflows in Databricks workspace with serverless compute.
Managed Identity Authentication (MANDATORY) -- Use managed identities for ALL Azure resources. Leverage Credentials feature for consolidation. MFA-compliant since October 2025.
Monitor Job Execution -- Track Databricks Job run IDs from ADF output, log parameters for auditability, set up alerts for failures, leverage built-in lineage.
Optimize Spark 3.3 (Data Flows) -- Enable AQE, use 4-8 partitions per core, broadcast joins for small dimensions, dynamic partition pruning.
Resources
Progressive Disclosure References
- Databricks Job Examples:
references/databricks-job-examples.md - Complete JSON for Job activity, linked services, pipeline, and Databricks workspace job definition
- Connector Examples:
references/connector-examples.md - Complete JSON for ServiceNow V2, PostgreSQL, Snowflake, Azure Storage MI, Mapping Data Flows, and Azure DevOps Server
1---2name: databricks-20253description: ADF + Databricks 2025 integration patterns. PROACTIVELY activate for: (1) Databricks Job activity in ADF, (2) DatabricksJob (preview) vs DatabricksNotebook activity, (3) ServiceNow V2 connector, (4) ADF managed identity authentication for Databricks, (5) Databricks serverless linked services, (6) Snowflake V2 connector, (7) Databricks job parameters and outputs, (8) MFA enforcement and authentication updates, (9) Unity Catalog integration, (10) Delta Live Tables orchestration from ADF. Provides: Databricks linked service templates (PAT, MSI, serverless), DatabricksJob activity examples, parameter passing recipes, and authentication migration guidance.4---5
6# Azure Data Factory Databricks Integration 2025
7
8## Databricks Job Activity (Recommended 2025)
9
10**CRITICAL UPDATE (2025):** The Databricks Job activity is now the **ONLY recommended method** for orchestrating Databricks in ADF. Microsoft strongly recommends migrating from legacy Notebook, Python, and JAR activities.
11
12### Quick Reference
13
14- **Activity type:** `DatabricksJob` (NOT `DatabricksSparkJob` or `DatabricksNotebook`)
15- **Parameter property:** `jobParameters` (NOT `parameters`)
16- **Linked service auth:** Managed Identity (`"authentication": "MSI"`) recommended
17- **Cluster config:** Do NOT specify cluster properties in linked service; the Databricks Job controls compute
18
19### Why Databricks Job Activity?
20
21| Feature | Notebook Activity (Legacy) | Job Activity (2025) |
22|---------|---------------------------|---------------------|
23| Compute | Must configure cluster in linked service | Serverless by default |
24| Workflow tasks | Single notebook | Multi-task DAGs (notebook, Python, SQL, DLT) |
25| Retry | ADF-level only | Job-level + task-level |
26| Repair runs | Not supported | Rerun failed tasks only |
27| Git integration | Limited | Full Databricks Git support + DABs |
28| Lineage | None | Built-in data lineage |
29| If/Else logic | Must use ADF control flow | Native If/Else task types |
30
31### Benefits Summary
32
331. **Serverless Execution** -- No cluster specification needed; automatic serverless compute with faster startup and lower costs
342. **Advanced Workflow Features** -- Run As, Task Values, Conditional Execution, AI/BI Tasks, Repair Runs, Notifications, Queuing
353. **Centralized Job Management** -- Jobs defined once in Databricks workspace; single source of truth with Git-backed versioning
364. **Cost Optimization** -- Serverless compute (pay only for execution), job clusters (auto-terminating), spot instance support
37
38For complete JSON examples of Job activity, linked service, and pipeline configurations, see `references/databricks-job-examples.md`.
39
40## Connectors and Enhancements (2025+)
41
42### ServiceNow V2 Connector (RECOMMENDED - V1 End of Support)
43
44**ServiceNow V1 connector is at End of Support. Migrate to V2 immediately.**
45
46| Feature | V1 | V2 |
47|---------|----|----|
48| Linked service type | `ServiceNow` | `ServiceNowV2` |
49| Source type | `ServiceNowSource` | `ServiceNowV2Source` |
50| Query builder | Custom | Aligns with ServiceNow condition builder |
51| Performance | Standard | Enhanced extraction |
52| OData support | No | Yes |
53
54**Migration steps:** Update linked service type to `ServiceNowV2`, update source type to `ServiceNowV2Source`, test queries in ServiceNow UI condition builder, adjust timeouts.
55
56### Enhanced PostgreSQL Connector
57
58Improved performance with 2025 SSL enhancements: `enableSsl: true`, `sslMode: "Require"`.
59
60### Enhanced Snowflake Connector
61
62Improved performance with KeyPair authentication support and Key Vault secret integration.
63
64### Managed Identity for Azure Storage
65
66New managed identity support for Azure Table Storage and Azure Files connectors (system-assigned and user-assigned).
67
68### Mapping Data Flows - Spark 3.3
69
70Spark 3.3 now powers Mapping Data Flows with 30% faster processing, Adaptive Query Execution (AQE), dynamic partition pruning, improved caching, and better column statistics.
71
72### Azure DevOps Server 2022 Support
73
74Git integration now supports on-premises Azure DevOps Server 2022 via the `hostName` property.
75
76For complete JSON examples of all connectors, see `references/connector-examples.md`.
77
78## Managed Identity 2025 Best Practices
79
80### User-Assigned vs System-Assigned
81
82| Scenario | Recommendation |
83|----------|---------------|
84| Single ADF, simple setup | System-assigned |
85| Multiple data factories | User-assigned (shared identity) |
86| Complex multi-environment | User-assigned |
87| Granular permission control | User-assigned |
88| Identity lifecycle independence | User-assigned |
89
90Use ADF's centralized **Credentials** feature to consolidate Microsoft Entra ID-based credentials across multiple linked services.
91
92### MFA Enforcement (Enforced Since October 2025)
93
94Azure MFA is mandatory for all interactive user logins. Impact on ADF:
95- Managed identities are **UNAFFECTED** -- no MFA required for service accounts
96- Service principals with certificate auth are the recommended alternative to secrets
97- All interactive user logins require MFA
98
99### Principle of Least Privilege
100
101| Resource | Source Role | Sink Role |
102|----------|-----------|-----------|
103| Storage Blob | `Storage Blob Data Reader` | `Storage Blob Data Contributor` |
104| SQL Database | `db_datareader` | `db_datareader` + `db_datawriter` |
105| Key Vault | `Get` secrets only | `Get` secrets only |
106
107For complete managed identity JSON examples, see `references/connector-examples.md`.
108
109## Best Practices (2025)
110
1111. **Use Databricks Job Activity (MANDATORY)** -- Stop using Notebook, Python, JAR activities. Define workflows in Databricks workspace with serverless compute.
112
1132. **Managed Identity Authentication (MANDATORY)** -- Use managed identities for ALL Azure resources. Leverage Credentials feature for consolidation. MFA-compliant since October 2025.
114
1153. **Monitor Job Execution** -- Track Databricks Job run IDs from ADF output, log parameters for auditability, set up alerts for failures, leverage built-in lineage.
116
1174. **Optimize Spark 3.3 (Data Flows)** -- Enable AQE, use 4-8 partitions per core, broadcast joins for small dimensions, dynamic partition pruning.
118
119## Resources
120
121- [Databricks Job Activity](https://learn.microsoft.com/azure/data-factory/transform-data-using-databricks-spark-job)
122- [ADF Connectors](https://learn.microsoft.com/azure/data-factory/connector-overview)
123- [Managed Identity Authentication](https://learn.microsoft.com/azure/data-factory/data-factory-service-identity)
124- [Mapping Data Flows](https://learn.microsoft.com/azure/data-factory/concepts-data-flow-overview)
125
126## Progressive Disclosure References
127
128- **Databricks Job Examples**: `references/databricks-job-examples.md` - Complete JSON for Job activity, linked services, pipeline, and Databricks workspace job definition
129- **Connector Examples**: `references/connector-examples.md` - Complete JSON for ServiceNow V2, PostgreSQL, Snowflake, Azure Storage MI, Mapping Data Flows, and Azure DevOps Server
josiahsiegel