Framework Deployment Options¶
The Lakeflow Framework can be deployed to a Databricks workspace in three ways. Choose the mode that matches how your team manages dependencies and configuration.
Mode |
How it works |
When to use |
|---|---|---|
Flat DAB deploy (default) |
The framework repository is cloned and deployed with
|
Default for all existing customers. No change to existing pipelines required. |
Wheel install |
|
Teams that manage Python dependencies via PyPI, a UC Volume, or an internal Artifactory feed. |
Wheel + local overlay |
Wheel installed, plus |
Teams that want pip-managed installs but need per-deployment config customization. |
Note
For wheel based deployments it is your responsibility to provide an entry-point notebook and a location for local config overrides and libraries/init scripts when moving away from the default flat deploy.
See also¶
Deploy framework from local machine — flat DAB deploy from your laptop
Installing the Framework as a Wheel — install
lakeflow-frameworkas a Python wheel