Skip to main content

Automate

Automation is key to building reliable, efficient data pipelines. Dagster offers several ways to automate pipeline execution to fit a variety of needs.

Consider these factors when selecting an automation method:

  • Pipeline structure: Are you working primarily with assets, ops, or a mix?
  • Timing requirements: Do you need regular updates or event-driven processing?
  • Data characteristics: Is your data partitioned? Do you need to update historical data?
  • System integration: Do you need to react to external events or systems?

Automation methods

MethodDescriptionBest forWorks with
SchedulesRun a selection of assets at specified times with a cron expressionRegular, time-based job runs and basic time-based automationAssets, Ops, Graphs
Declarative automationA framework that allows you to set automation conditions on assets and asset checksAsset-centric, condition-based updatesAssets only
SensorsTrigger runs based on events or conditions that you define, like the arrival of a new file or a change to an external systemEvent-driven automationAssets, Ops, Graphs
Asset sensorsTrigger jobs when specified assets are materialized, allowing you to create dependencies between jobs or code locations.Cross-job/location asset dependenciesAssets only
GraphQL triggersTrigger materializations and jobs from the GraphQL endpointEvent triggers from external systemsAssets, Ops, Jobs