Dagster Block

The Dagster block integrates Dagster data orchestration into your workflows. Use it to launch runs, monitor their status, retrieve run logs, list recent executions, terminate stuck runs, and enumerate available jobs and schedules in a Dagster deployment.

Overview

When to Use

• Launch a Dagster job run with a custom run configuration as part of an automated pipeline.
• Poll or inspect the status and metadata of an existing run by its ID.
• Pull structured execution event logs to diagnose pipeline failures or pass results to downstream blocks.
• List recent runs, filtered by job name or status, to find in-progress or failed executions.
• Terminate a run that is stuck, timed out, or no longer needed.
• Enumerate all jobs across repositories or list all schedules (with their cron expressions and status) in a Dagster instance.

Configuration

Operation

Required. Choose which Dagster action to perform. Every other field is shown or hidden based on this selection.

Dagster URL (baseUrl)

Required for all operations. The base URL of your Dagster deployment, for example https://dagster.example.com. The block appends /graphql to this URL for all API calls.

To keep the URL out of plain text, store it in an environment variable and reference it as {{DAGSTER_BASE_URL}}.

Repository Location (repositoryLocationName)

Shown for: dagster_launch_run, dagster_list_jobs.

The code-location (repository location) name as configured in your Dagster workspace, for example my_repo_location. Required when launching a run or listing jobs.

Repository Name (repositoryName)

Shown for: dagster_launch_run, dagster_list_jobs.

The repository name within the code location, for example __repository__. Required when launching a run or listing jobs.

Job Name (jobName)

Shown for: dagster_launch_run.

The name of the Dagster job (pipeline) to launch, for example my_pipeline. Required for Launch Run.

Run Config (JSON) (runConfigJson)

Shown for: dagster_launch_run.

Optional run configuration passed to the job as a JSON object, for example:

{"ops": {"my_op": {"config": {"param": "value"}}}}

Leave blank to use the job's default configuration.

Run ID (runId)

Shown for: dagster_get_run, dagster_get_run_logs, dagster_terminate_run.

The globally unique identifier of an existing run. Required for Get Run, Get Run Logs, and Terminate Run. You can pass the output of a Launch Run step using {{dagster_1.runId}}.

Inputs & Outputs

• Inputs:

  • operation (string) — Operation to perform (one of the operation IDs above)
  • baseUrl (string) — Dagster URL
  • repositoryLocationName (string) — Repository location name (Launch Run, List Jobs, List Schedules)
  • repositoryName (string) — Repository name (Launch Run, List Jobs, List Schedules)
  • jobName (string) — Job name (Launch Run)
  • runConfigJson (json) — Run configuration as a JSON object (Launch Run, optional)
  • runId (string) — Run ID (Get Run, Get Run Logs, Terminate Run)

• Outputs (block-level, operation-dependent):

  • runId (string) — Run ID of the launched or terminated run
  • status (string) — Run status (e.g. SUCCESS, FAILURE, STARTED, TERMINATED)
  • logs (json) — Run logs

  Detailed per-operation outputs from the underlying tools are documented in the Tools section below.

Tools

Dagster Launch Run (dagster_launch_run) — Launches a new Dagster job run via GraphQL mutation. Accepts repositoryLocationName, repositoryName, jobName, an optional runConfigJson (JSON string), and optional tags (JSON array of {key, value} objects). Returns runId (string) — the globally unique ID of the launched run. Optionally accepts apiKey (Dagster+ API token; omit for OSS).

Dagster Get Run (dagster_get_run) — Retrieves the status and full metadata of a run by ID. Returns runId (string), jobName (string, optional), status (string), startTime (number — Unix timestamp, optional), endTime (number — Unix timestamp, optional), runConfigYaml (string, optional), and tags (json, optional).

Dagster Get Run Logs (dagster_get_run_logs) — Fetches execution event logs for a run. Supports pagination via afterCursor and a limit parameter. Returns events (json — array of log event objects each with type, message, timestamp, level, stepKey, eventType), cursor (string, optional — pass to the next call for pagination), and hasMore (boolean).

Dagster List Runs (dagster_list_runs) — Lists recent runs, optionally filtered by jobName (string) and/or statuses (comma-separated string such as "SUCCESS,FAILURE"). Accepts a limit parameter (default 20). Returns runs (json — array of objects each with runId, jobName, status, tags, startTime, endTime).

Dagster Terminate Run (dagster_terminate_run) — Terminates an in-progress run. Returns runId (string — the run that was terminated) and status (string — TERMINATED).

Dagster List Jobs (dagster_list_jobs) — Lists all jobs across all repositories in the Dagster instance. Returns jobs (json — array of objects each with name and repositoryName).

Dagster List Schedules (dagster_list_schedules) — Lists all schedules defined in a repository. Requires repositoryLocationName and repositoryName. Accepts an optional scheduleStatus filter (RUNNING or STOPPED). Returns schedules (json — array of objects each with name, cronSchedule, jobName, status, id, description, executionTimezone).

All tools accept an optional apiKey param (Dagster+ API token). Leave it blank for open-source Dagster deployments; pass {{DAGSTER_API_KEY}} for Dagster Cloud.

YAML Example

# Launch a Dagster job run and pass its ID to the next block
dagster_1:
  type: dagster
  name: "Launch ETL Pipeline"
  inputs:
    operation: "dagster_launch_run"
    baseUrl: "{{DAGSTER_BASE_URL}}"
    repositoryLocationName: "my_repo_location"
    repositoryName: "__repository__"
    jobName: "my_pipeline"
    runConfigJson: '{"ops": {"my_op": {"config": {"param": "value"}}}}'
  connections:
    outgoing:
      - target: next-block-id