dagster

Start development server

$ dagster dev

$ dagster dev -m [my_project]

$ dagster-daemon run

$ dagster-webserver

$ dagster job execute -m [my_project] -j [job_name]

$ dagster job list -m [my_project]

$ dagster project scaffold --name [my-project]

$ dagster definitions validate -m [my_project]

dagster command [options]
dagster-daemon run [options]
dagster-webserver [options]

dagster is the CLI for Dagster, a data orchestration platform built around software-defined assets. It manages the development environment, job execution, and infrastructure.
dagster dev starts both the webserver (UI) and daemon (schedules, sensors) for local development. In production, run dagster-webserver and dagster-daemon separately.
Software-defined assets are the core abstraction—functions that produce data assets with dependencies. Assets form a DAG that Dagster materializes. Jobs group assets for execution.
Schedules trigger jobs on cron patterns; sensors trigger based on external events. Both require the daemon process to run.
The webserver provides a UI showing asset lineage, run history, logs, and metrics. The asset graph visualizes data dependencies.

-m, --module name

Python module containing definitions.

Python file containing definitions.

Job name.

Webserver port. Default: 3000.

Webserver host. Default: localhost.

Workspace YAML file.

Working directory for code.

dagster.yaml

Instance configuration file controlling storage, compute, and run settings.

Workspace configuration defining code locations and repositories.

dev

Start development server (webserver + daemon).

Manage and run jobs.

Manage software-defined assets.

Manage schedules.

Manage sensors.

Create new projects.

Validate code definitions.

Manage Dagster instance.

Manage pipeline runs.

Daemon is required for schedules and sensors. Asset materialization tracks state in instance storage. Production requires PostgreSQL for run storage. Some features require Dagster+ (cloud).

Dagster was created by Elementl, founded by Nick Schrock (co-creator of GraphQL) in 2018. The project introduced the concept of software-defined assets as an improvement over task-oriented workflows. Version 1.0 released in 2022. Dagster emphasizes developer experience with type checking, testing utilities, and local development. The company offers Dagster+ for managed cloud orchestration.

prefect(1), airflow(1), dbt(1), argo(1)