Dagit Web UI

1.13.0 · active · verified Wed Apr 15

Dagit is the web-based user interface for Dagster, providing observability, operational control, and debugging capabilities for data pipelines. It allows users to visualize assets, jobs, runs, schedules, and sensors defined within a Dagster project. Dagit's versioning is tightly coupled with the core Dagster library, typically released in sync with major and minor Dagster updates.

Warnings

breaking Dagit (and the underlying `dagster-webserver` package) versions must closely match the `dagster` core library version. Mismatched versions (e.g., `dagster==1.12.0` and `dagit==1.13.0`) can lead to runtime errors or unexpected UI behavior.
Fix: Always install `dagit` and `dagster` with matching major and minor versions (e.g., `pip install 'dagster==1.13.*' 'dagit==1.13.*'`).
gotcha Dagit is primarily a CLI executable (`dagit`) used to launch the Dagster UI. It is not designed to be imported as a Python module within user-defined Dagster code (e.g., `import dagit`).
Fix: Interact with Dagit via the command line. Define your Dagster assets, ops, and jobs using the `dagster` library, and then use `dagit -f path/to/your_definitions.py` to view them in the UI.
breaking Since `dagster` version 1.12.18, `psycopg2-binary` is no longer a transitive dependency of `dagster-postgres`. If your Dagster project utilizes `dagster-postgres` and relies on `psycopg2-binary`, you must explicitly install it.
Fix: If you use `dagster-postgres`, explicitly add `pip install psycopg2-binary` to your project's dependencies.
gotcha Running `dagit` without specifying any Dagster definitions (e.g., via `-f`, `-m`, or a `workspace.yaml` file) will result in an empty UI. Dagit needs to know where to find your code.
Fix: Always provide a target for Dagit to load definitions from. For local development, use `dagit -f your_definitions.py` or `dagit -m your_package`.

Install

pip install dagit Install Dagit

Imports

N/A
```
dagit is a CLI tool, not a Python library for direct import.
```
The `dagit` package primarily provides the `dagit` command-line executable, which launches the Dagster UI. Users typically run `dagit` from the terminal rather than importing symbols from it in Python code.

Quickstart

To start Dagit, first define some Dagster assets or jobs in a Python file (e.g., `my_repo.py`). Then, launch Dagit from your terminal, pointing it to your definitions file. This will open the Dagster UI in your browser.

from dagster import asset, Definitions, job

@asset
def hello_asset():
    return "Hello, Dagster World!"

@job
def my_simple_job():
    hello_asset()

defs = Definitions(assets=[hello_asset], jobs=[my_simple_job])

# Save this as `my_repo.py` then run `dagit -f my_repo.py` in your terminal.

view raw JSON →