Deploying Dagster

Getting started with Dagster is as easy as running:

pip install dagit

Dagster supports a wide variety of more sophisticated deployment strategies for production.

Typically, you'll want to deploy Dagit somewhere, configure the way that it stores information and launches runs, and make sure your pipelines are configured to run appropriately for your environment.

If you're using the Celery engine, you'll want to run worker processes to parallelize execution.

Alternatively, you may want to deploy Dagster pipelines into a completely different scheduling and execution environment, like Apache Airflow.

No matter how you're planning to deploy Dagster, you should first understand its default behavior.

You should also be familiar with configuring the Dagster instance, which organizes all of the information specific to a particular installation or deployment of Dagster, using dagster.yaml.

And you should be aware of the per-pipeline run configuration you'll need for your pipeline runs to take advantage of the environment in which they're deployed.