Databricks Job Viewer — VS Code Extension

Test version — This extension was built to answer my own day-to-day needs working with Databricks LakeFlow YAML job definitions. It is shared as-is. If it fits your workflow too, feel free to fork it and adapt it.

Author: Skander Boudawara — skander.education@proton.me

Open source — Fork it, improve it, make it yours.

Part of the code was written with the assistance of AI.

What does this extension do?

When you work with Databricks Declarative Automation Bundles, your jobs are defined as YAML files. Reading and editing those files in a text editor quickly becomes painful as the number of tasks grows.

This extension gives you a visual DAG (Directed Acyclic Graph) panel directly inside VS Code that lets you see the full structure of a job at a glance — and edit it without touching the YAML by hand.

Screenshots

Databricks UI (reference)

Extension UI

Features

Sidebar — Job Browser

• A searchable Job Workflows panel in the VS Code activity bar lists all jobs found under resources/app_jobs/**/*.yml and resources/etl_jobs/**/*.yml.
• Jobs are grouped by category and sub-category (folder structure), with collapsible sections.
• Fuzzy search bar at the top of the panel — type any substring or fuzzy pattern (e.g. "kay" matches "kaylani_etl" or "k_async_yearly"). Matched characters are highlighted in the results.
• Each job entry shows the task count and an Open YAML (↗) button.
• Collapse state is persisted across reloads.
• Refresh button in the panel title reloads all jobs from disk.

DAG Canvas — Job Visualization

• Opens a full interactive DAG panel when you click a job.
• Tasks are laid out automatically using a topological sort (left-to-right by dependency level).
• Each task node shows its type (Notebook, Spark Python, Python Wheel, SQL, Run Job, Condition), name, and notebook path (if applicable).
• Condition tasks are rendered as diamonds with True/False output ports.
• Color-coded edges — grey for unconditional, green for true, red for false outcomes.
• Outcome labels on edges.
• Pan (drag), zoom (scroll wheel or buttons), and Fit to view button.

Task Interaction

• Click a task to open its detail panel in the bottom bar.
• Double-click a notebook task to open the notebook file in the editor.
• Drag a task to reposition it on the canvas.

Port-based Dependency Editing

• Each task node has port circles (In on the left, Out on the right; Condition tasks have True/False output ports).
• Drag from an output port to an input port to create a new dependency edge.
• Hover an edge to reveal a × button — click it to delete that dependency.

Task Panel (Bottom Bar)

Clicking a task opens a detail panel at the bottom of the canvas with the following editable sections:

• Task name — rename inline (updates all depends_on references automatically on save).
• Run if — selector for ALL_SUCCESS, AT_LEAST_ONE_SUCCESS, NONE_FAILED, etc.
• Upstream Dependencies — interactive list:
  • Click a dependency to navigate to that task.
  • − button to remove a dependency.
  • Add upstream task dropdown + outcome selector (any / true / false) to add a new dependency.
• Condition (read-only display for condition tasks).
• Compute — job cluster and environment selectors, max retries, retry interval.
• Task Parameters — editable key/value pairs.
• Libraries — view current libraries with a Modify Libraries button to open the library editor.
• Remove task — red button at the bottom that immediately removes the task from the YAML (no Save needed). All depends_on references to that task in other tasks are cleaned up automatically.

Add New Task

• A blue + button (circle, bottom-right of the canvas) opens the New Task form in the bottom bar.
• Choose the task type: Notebook, Run Job (reference another job from the workspace), or Condition.
• Fill in the type-specific fields (notebook path / job name / condition operator + values).
• Click Add to DAG — the task is immediately written to the YAML file and appears on the canvas. No Save button needed.
• The new task can then be connected to others via port dragging.

Job Panel (Right Sidebar)

The right sidebar shows job-level properties when no task is selected:

• Job name — editable directly in the topbar.
• Description — editable text field.
• Trigger — dropdown to change trigger type (Manual, Scheduled/Cron, Table Update, File Arrival) with type-specific fields:
  • Scheduled: cron expression + timezone, or interval + unit.
  • Table Update: "Trigger when" condition (Any / All tables updated) + table list.
  • File Arrival: URL field.
• Timeout and Max concurrent runs.
• Queue enabled toggle.
• Health rules — editable threshold values.
• Job Parameters — editable default values.
• Job Clusters — view, edit, add, and remove job clusters with full configuration (node type, Spark version, worker mode: fixed / autoscale / single node, init scripts, Spark conf, etc.).
• Environments — view, edit, add, and remove serverless environments with dependency lists.
• Permissions — list of users/groups/service principals with their permission levels.
• Email Notifications display.

Editing & Saving

• All edits are tracked as pending changes (highlighted Save button).
• A Revert button cancels all pending changes and restores the last saved state.
• The Save button writes all pending changes to the YAML file. After saving, the panel reloads from the updated file.
• Task additions and removals are saved immediately without needing the Save button.

Installation

Download the .vsix file from the Releases page and install it in VS Code:

Extensions panel → ··· menu → Install from VSIX…

Or via the CLI:

code --install-extension databricks-job-viewer-<version>.vsix

The extension activates automatically when your workspace contains resources/**/*.yml or databricks.yml files.

Requirements

• VS Code ^1.85.0
• A Databricks bundle project with job YAML files under resources/app_jobs/ or resources/etl_jobs/

License

MIT — do whatever you want with it.