The easiest way to launch the UI from the command line during local development is to run:
dagster dev
This command launches both the Dagster webserver (which serves the UI) and the Dagster daemon, allowing you to start a full local deployment of Dagster from the command line.
The command will print out the URL you can access the UI from in the browser, usually on port 3000.
When invoked, the UI will fetch definitions - such as assets, jobs, schedules, sensors, and resources - from a Definitions object in a Python module or package or the code locations configured in an open source deployment's workspace files. Refer to the Code location documentation for more info.
You can also launch the webserver by itself from the command line by running:
dagster-webserver
Note that several Dagster features, like schedules and sensors, require the Dagster daemon to be running in order to function.
The Runs page lists all job runs, which can be filtered by job name, run ID, execution status, or tag. Click a run ID to open the Run details page and view details for that run:
The Run details page contains details about a single run, including timing information, errors, and logs.
The upper left pane contains a Gantt chart, indicating how long each op took to execute. The bottom pane displays filterable events and logs emitted during the course of execution:
From the Run details page, you can also re-execute a run using the same configuration by clicking the Re-execute button:
There are several re-execution options:
All Steps: Re-execute the run from scratch.
Selected Steps: Re-execute the selected steps.
From Selected: Re-execute the steps downstream from the selected steps
From Failure: Retry the run, skipping steps completed successfully. This is only enabled when the run has failed.
Related runs (e.g., runs created by re-executing the same previous run) are grouped together in the right pane for easy reference.
The Deployment overview page includes information about the status of the code locations in your deployment, daemon (Open Source) or agent (Cloud) health, schedules, sensors, and configuration details: