Deploy a Databricks app

After you create and develop your Azure Databricks app, deploy it to make it accessible in the Azure Databricks workspace. Deployment builds your app, installs dependencies, and runs it using the configuration defined in your project files. You can deploy apps using the Azure Databricks UI or the Azure Databricks CLI.

Deployment logic

Databricks Apps supports deploying applications that use either Python, Node.js, or a combination of both. This allows for flexible architectures, such as a Node.js frontend with a Python backend.

During deployment, the build process checks for a package.json file at the root of your app directory to determine whether Node.js is used. If present, it includes Node-specific build steps alongside Python steps. The deployment logic follows this pattern:

If package.json is present:

1. Run npm install
2. Run pip install -r requirements.txt (if it exists)
3. Run npm run build (if a build script is defined in package.json)
4. Run the command specified in app.yaml, or npm run start if no command is specified

If package.json is not present:

1. Run pip install -r requirements.txt (if it exists)
2. Run the command specified in app.yaml, or python <my-app>.py if no command is specified

Prepare for deployment

Before you deploy your app, verify that your project includes the necessary components:

• Main script - Your entry point file, such as app.py or app.js.
• Optional app.yaml file - If your app requires a custom command or environment variables, include an app.yaml file to configure execution. See Configure Databricks app execution with app.yaml.
• Dependencies - Make sure all dependencies are available. See Manage dependencies for a Databricks app.
• Secrets or environment values - If you use the env section in app.yaml, verify that the referenced secrets or external sources are correctly configured and accessible. See Add resources to a Databricks app.

In addition, make sure the app service principal has access to the source code folder.

Deploy the app

Databricks UI

To deploy an app from the Azure Databricks UI:

1. Upload the app files to your Azure Databricks workspace. For instructions, see Import a file.
2. Click Compute in the sidebar.
3. Go to the Apps tab and click the link to your app in the Name column.
4. Click Deploy and select the folder in your workspace where you uploaded the app files.
5. Click Select, then Deploy.
6. Review the configuration and click Deploy.

Databricks CLI

To deploy an app using the CLI:

1. Open a terminal and navigate to the directory that contains your app files.

2. Upload your app files to the Azure Databricks workspace using the sync command. Replace the path with the workspace location where you want to upload the files.

   databricks sync --watch . /Workspace/Users/my-email@org.com/my-app
   

   The --watch flag keeps the sync process running and automatically uploads changes when you modify files locally. To exclude specific files or directories from syncing, add them to a .gitignore file in your local app directory. Common files to exclude are node_modules/, .env, __pycache__/, .DS_Store, and any large data files or build artifacts.

3. Verify the upload by viewing the files in your workspace. Click Workspace in the sidebar and navigate to the directory you created for your app.

4. Deploy the app by running the following command. Replace the app name and source code path with your values.

   databricks apps deploy my-app-name \
      --source-code-path /Workspace/Users/my-email@org.com/my-app
   

   The CLI displays deployment progress and confirms when the app is running.

Once deployed, Azure Databricks starts your app based on the defined command in your app.yaml file or defaults to running python app.py.

Post-deployment behavior

After deployment completes, Azure Databricks starts your app in a managed environment. The app details page shows the current status and provides access to logs, deployment history, and environment information.

To view the deployed app's output, click the app link.

Go to the Logs tab for debugging and runtime monitoring. See Logging and Monitoring for Databricks Apps.

Update or redeploy the app

You might want to update or re-deploy your app if you made changes to the source code, updated the app.yaml file, or need to deploy from a different source code path. Re-deploying applies your latest updates without needing to recreate the app from scratch.

To make changes to a deployed app:

1. Update your app code or app.yaml configuration.
2. Return to the Apps page in the Azure Databricks workspace.
3. Select your app and choose Deploy again. If the source code path changed, click the arrow next to Deploy and select Deploy using different source code path.

Troubleshoot deployment issues

If your app fails to deploy or doesn't run as expected, try the following troubleshooting steps:

• Check logs - View error messages or runtime output in the logs panel.
• Validate app.yaml - Make sure the syntax is correct and supported settings are used.
• Verify secrets and environment variables - Make sure all references in the env section resolve properly.
• Confirm dependencies - Make sure all required packages are included or installed.

Next steps

• Logging and Monitoring for Databricks Apps
• Configure permissions for a Databricks app
• Configure Databricks app execution with app.yaml