![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(268.8, 56.00000000000001%, 36%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EJC%3C/text%3E%3C/svg%3E)
Azure Databricks
An Apache Spark-based analytics platform optimized for Azure.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(268.8, 8%, 36%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EST%3C/text%3E%3C/svg%3E)
Hi @Janice Chi
Thanks for outling your scenario. Given your move toward a multi-environment CI/CD model for healthcare-grade pipelines, it's great to see you're evaluating Databricks Asset Bundles (DAB) alongside your existing Terraform-based infrastructure.
Is DAB the recommended approach for multi-environment Databricks deployments?
Yes – Databricks Asset Bundles (DAB) are now the recommended and supported approach by Databricks for modular, code-driven deployments of notebooks, jobs, workflows, and environment-specific configurations.
They integrate well with Azure DevOps, GitHub Actions, and other CI/CD tools, making them suitable for structured promotion across Dev → QA → PreProd → Prod.
How does DAB compare with Terraform?
| Area | Databricks Asset Bundles (DAB) | Terraform for Databricks |
|---|---|---|
| Primary Purpose | Application deployment (jobs, workflows, code) | Infrastructure provisioning (clusters, ACLs, etc.) |
| CI/CD Integration | Built-in support (e.g., bundle deploy, bundle run) |
Requires external scripting |
| Environment Overrides | Native support via targets: in bundle.yml |
Requires variable files and conditional logic |
| Secrets/Mounts | Not supported (must be handled separately) | Can provision secrets, scopes, mounts |
| Unity Catalog Support | Fully supported in latest DAB versions | Supported |
| Versioning | Git-based, declarative (via repo + DAB CLI) | Declarative (via .tf files) |
| Deployment to Repos | Supports linking to DBR Repos as source | Terraform supports repos via API or init scripts |
DAB and Terraform are complementary - not competing. Most customers use:
- Terraform to provision infrastructure
- DAB to deploy jobs, notebooks, workflows, and environment-specific configurations
How does DAB work with DBR Repos?
- DAB can deploy to Databricks Repos directly, allowing developers to version notebooks in Git and sync them with the workspace.
- You can also define repos as dependencies or sources in
bundle.yml. - Supports workflows that run code from repos, not just workspace paths.
Healthcare Considerations & Best Practices
Given the regulatory and audit requirements:
- Keep secrets (e.g., connection strings, SHIR names) outside the bundle, injected at deploy/runtime via:
- Azure Key Vault integration
- Pipeline environment variables
run_id,env,git_commitfor audit traceability. - Use target overrides in
bundle.ymlto:- Point to different SQL endpoints
- Change notebook parameters
- Control retry logic per environment
Known DAB Limitations
Doesn’t manage:
- Secrets (must use Terraform or CLI)
- Mount points
- Unity Catalog permissions (use Terraform or REST APIs)
No direct support for Delta Live Tables (DLT) yet
Testing for large bundles (>100 jobs) may need careful orchestration
Useful Docs:
Hope this helps. If this answers your query, do click Accept Answer and Yes for was this answer helpful. And, if you have any further query do let us know.