Bundle configuration

Bundle support turns on whenever a databricks.yml file is present in the project root. Pass --no-bundle to lhp generate or lhp validate to disable it. LHP never reads or modifies databricks.yml; it only detects the file’s presence and writes pipeline resource YAML under resources/lhp/.

See also

How-to guides: Validate and ship pipelines in CI, Deploy and run.

Activation flags

Flag	Command	Behavior
--no-bundle	generate, validate	Skip bundle detection and resource sync even when databricks.yml exists.
-pc, --pipeline-config	generate, validate	Path (relative to project root) to the pipeline_config.yaml file. Required when bundle support is enabled.
--force	generate	Deprecated no-op. Every run fully regenerates; the flag is hidden and ignored.

When bundle support is enabled and -pc / --pipeline-config is omitted, both generate and validate fail up-front with LHP-CFG-023 before any files are written. Supply the flag, or pass --no-bundle to skip bundle resource generation entirely.

databricks.yml

User-owned; LHP reads only its presence. These fields govern how the Databricks CLI picks up LHP-generated resources.

Field	Type	Description
bundle.name	string	Bundle name used by the Databricks CLI for deploy paths.
include	list[string]	Resource-file globs. Must cover resources/lhp/*.yml so the Databricks CLI loads LHP-generated pipeline resources.
targets	mapping	One entry per deployment target. Each target name must match an LHP environment — a substitutions/<env>.yaml file with the same name.

bundle:
  name: my_project

include:
  - resources/*.yml
  - resources/lhp/*.yml

targets:
  dev:
    mode: development
    default: true
    workspace:
      host: <databricks_host>
  prod:
    mode: production
    workspace:
      host: <databricks_host>

config/pipeline_config.yaml

A multi-document YAML file. The first document holds project_defaults; each subsequent document targets one or more pipelines via a pipeline key (a single name or a list). Pass its path with -pc / --pipeline-config.

project_defaults:
  <shared keys>
---
pipeline: <name>          # or [name1, name2]
<per-pipeline keys>

Top-level keys

Valid under both project_defaults and a per-pipeline document.

Key	Type	Required	Default	Description
catalog	string	Yes	—	Unity Catalog name. Must be set with schema (both or neither) and non-empty after substitution. Supports ${token}.
schema	string	Yes	—	Schema name. Must be set with catalog. Supports ${token}.
serverless	bool	No	true	Pipeline compute mode.
edition	string	No	ADVANCED	One of CORE, PRO, ADVANCED. Ignored when serverless: true.
channel	string	No	CURRENT	One of CURRENT, PREVIEW.
continuous	bool	No	false	Streaming/continuous pipeline mode.
packaging	string	No	source	One of source, wheel. Selects how generated code ships (see Wheel packaging). Consumed by LHP; never written to the resource YAML.

Any other top-level key (clusters, configuration, notifications, tags, event_log, environment, permissions, photon, and any Databricks Pipelines API field such as run_as) is passed through verbatim into the generated pipeline resource.

Merge precedence (lowest to highest): built-in defaults → project_defaults → per-pipeline document. Nested mappings are deep-merged; lists are replaced wholesale. Token substitution from substitutions/<env>.yaml applies to every field.

project_defaults:
  catalog: "${catalog}"
  schema: "${bronze_schema}"
  serverless: true

---
pipeline: bronze_load
packaging: wheel

Wheel packaging

Set packaging: wheel on a pipeline (or under project_defaults) to ship its generated code as a built wheel instead of source files. Wheel mode requires a wheel.artifact_volume in lhp.yaml.

Key (lhp.yaml)	Type	Description
wheel.artifact_volume	string	Unity Catalog volume path where built wheels are uploaded. After substitution it must start with /Volumes/; otherwise wheel packaging fails with LHP-CFG-061. Supports ${token}.

wheel:
  artifact_volume: "/Volumes/${catalog}/artifacts/wheels"

Generated output

Each generated pipeline produces one resource file resources/lhp/<pipeline_name>.pipeline.yml carrying a # Generated by LakehousePlumber header. Every lhp generate run wipes resources/lhp/ and rewrites it — manual edits are lost.

Path	Contents
resources/lhp/<pipeline_name>.pipeline.yml	Pipeline resource; declares a <pipeline_name>_pipeline resource key.
resources/lhp/_wheels.bundle.yml	Written only when at least one pipeline uses packaging: wheel; points the bundle’s artifact upload path at wheel.artifact_volume.