Tags#

Over time, your Dagster project may grow to include a large number of Dagster definitions. Using tags to label assets, ops, jobs, and job runs can make them easier to find.

Benefits#

Using tags helps you:

• Organize definitions in your Dagster project
• Improve filtering in the Dagster UI, making it easier to focus on specific items
• Customize run execution, including run concurrency and priority

How it works#

Tags are key/value pairs of strings that can be applied to assets, ops, jobs, and job runs using the tags argument. For example:

@asset(tags={"domain": "marketing", "pii": "true"})
def leads(): ...

Tags are helpful when you need to find specific items in your Dagster project. For example, tags can be used as a filter when looking at the Asset catalog in the Dagster UI, allowing you to only see specific assets.

Tag keys#

Valid tag keys have two segments: an optional prefix and name, separated by a slash (/). Prefixes are typically expected to be Python package names. For example: dagster/priority

Prefixes and name segments must each:

• Be 63 characters or less
• Contain only alphanumeric characters, dashes (-), underscores (_), and dots (.)

Tag values#

Tag values must:

• Be 63 characters or less
• Contain only alphanumeric characters, dashes (-), underscores (_), and dots (.)
• Be a string or JSON that is serializable to a string

Specifying tags#

Assets#

Tags are useful for organizing assets non-hierarchical ways. Unlike asset groups, where each asset belongs to a single group, each asset can have many tags. For example:

@asset(tags={"domain": "marketing", "pii": "true"})
def leads(): ...

Other APIs that define assets can also accept tag arguments, such as AssetSpec, AssetOut, AssetOut, and @graph_asset.

Jobs#

Adding tags to a job will attach them to every run executed by the job.

asset_job = define_asset_job(
    name="marketing_job", selection="*", tags={"dagster/max_runtime": 10}
)

@job(tags={"domain": "marketing"})
def email_job():
    send_email()

my_second_job_with_tags = email_graph.to_job(tags={"domain": "marketing"})

System run tags#

In some cases, Dagster will automatically add tags to runs:

Customizing run execution#

While tags are primarily used for labeling and organization, some run execution features are controlled using run tags:

• Customizing Kubernetes config
• Specifying Celery config
• Setting concurrency limits when using the QueuedRunCoordinator
• Setting the priority of different runs

Viewing tags in the Dagster UI#

View tags attached to runs by navigating to the Runs page in the UI:

When executing a job, use the Launchpad to add tags to the run:

Related#