Skip to main content
Last updated on

Define environment variables in a Databricks app

Databricks automatically sets certain environment variables in the app runtime environment. These variables provide essential information about the app and workspace, and are accessible to all Databricks apps by default. For a list of default variables, see Databricks Apps environment.

If your app requires additional environment variables, define them in the app.yaml configuration file in the env section. Each variable requires a name and a value. Variables can use a hardcoded value or reference an external source.

For example:

YAML
env:
  - name: LOG_LEVEL
    value: 'debug'

Only hardcode values when they're static, non-sensitive, and consistent across environments. Examples include value: "true" for feature toggles, value: "us-west" for fixed regions, or value: "UTC" for default timezones.

important

To keep your app secure and portable, never reference secret keys or other sensitive values directly in your app configuration. For example, avoid embedding secret values in the value field of an environment variable or directly in your source code. Instead, use the valueFrom field to securely reference secrets and other managed resources defined in your resources block. This ensures secrets are retrieved from Databricks at runtime and are never exposed in plaintext in your configuration files.

Use environment variables to access resources​

If you define app resources, such as SQL warehouses or secrets, reference these resources in the env section of your app.yaml file using the valueFrom field. This connects environment variables in your app to the resource keys defined in resources.

Example app.yaml snippet:

YAML
env:
  - name: WAREHOUSE_ID
    valueFrom: sql_warehouse

  - name: SECRET_KEY
    valueFrom: secret

Then, in your app code, access them as environment variables:

Python
import os

warehouse_id = os.getenv("WAREHOUSE_ID")
secret_value = os.getenv("SECRET_KEY")

valueFrom reference​

The following table shows the value that valueFrom resolves to for each resource type:

Resource type

Resolved value

Example

Databricks app

App name

my-app

Genie space

Space ID

01ef1fa2b3c45678

Lakeflow job

Job ID

123456789

MLflow experiment

Experiment ID

456789012

Model serving endpoint

Endpoint name

my-serving-endpoint

Secret

Decrypted secret value

(the secret value)

SQL warehouse

Warehouse ID

a1b2c3d4e5f67890

Unity Catalog connection

Connection name

my_connection

Unity Catalog table

Table full name

catalog.schema.table

Unity Catalog volume

Volume path

/Volumes/catalog/schema/volume

User-defined function

Function full name

catalog.schema.my_function

Vector search index

Index full name

catalog.schema.my_index

Next steps​