Skip to main content

Running Your Pipelines in Dagster

Quick Start: Run Your First Pipeline

  1. Navigate to the Assets page
  2. Click Materialize all in the top right
  3. Confirm the run
  4. Monitor progress in the Runs page

Method 2: Run Specific Assets

  1. Go to Assets page
  2. Select assets by checking boxes
  3. Click Materialize selected
  4. Useful for testing individual components

Method 3: Launch a Job

  1. Navigate to Jobs page
  2. Find your job (e.g., meltano_dbt_pipeline)
  3. Click Launch Run
  4. Configure if needed (usually not required)

Understanding Your Pipeline Structure

Typical Pipeline Flow

Meltano Extractors → Raw Data → dbt Models → Analytics Tables
(Assets) (Assets) (Assets) (Assets)

Asset Dependencies

  • Meltano assets must complete before dbt assets
  • Dependencies are automatically managed
  • Failed upstream assets prevent downstream execution

Monitoring Your Runs

Run Status Indicators

  • 🟢 Success: Asset materialized successfully
  • 🔵 In Progress: Currently executing
  • 🔴 Failed: Error occurred (check logs)
  • Queued: Waiting to start

Viewing Run Details

  1. Click on any run in the Runs page
  2. See:
    • Timeline of execution
    • Logs for each step
    • Asset materializations
    • Error messages (if any)

Understanding Logs

Meltano Logs

2024-01-15 10:30:45 [info] Running extract/load for tap-salesforce
2024-01-15 10:30:46 [info] Updated state: 2024-01-15T10:30:46
2024-01-15 10:31:15 [info] Extracted 1,543 records

dbt Logs

10:31:20  Running dbt...
10:31:25 Found 23 models, 15 tests
10:31:30 Completed successfully

Validating Your Migration

Checklist for First Run

  • All Meltano extractors complete successfully
  • Data lands in expected raw tables
  • dbt models build without errors
  • Row counts match expectations
  • Schedules show correct next run time

Comparing with Arch

Check Data Freshness

  1. Go to Assets page
  2. Look at "Last Materialized" timestamps
  3. Compare with last Arch run times

Verify Row Counts

-- Run in your data warehouse
SELECT COUNT(*) FROM raw.salesforce_accounts;
SELECT COUNT(*) FROM analytics.customer_summary;

Handling Common Scenarios

Rerunning Failed Assets

  1. Click on the failed asset
  2. Review error in logs
  3. Click Re-materialize after fixing issue

Running Historical Backfills

  1. Select assets needing backfill
  2. Click Materialize selected
  3. Choose "Custom Run Config"
  4. Specify date range for Meltano

Partial Pipeline Runs

Run only dbt models (skip extraction):

  1. Select only dbt assets
  2. Materialize selected
  3. Useful for model development

State Management

Meltano State

  • Automatically managed by platform
  • Tracks last successful extraction
  • Prevents duplicate data
  • No manual intervention needed

Viewing State

  1. Click on Meltano asset
  2. Check "Metadata" tab
  3. See last bookmark value

Scheduling Runs

View Existing Schedules

  1. Go to Schedules page
  2. See all migrated schedules
  3. Check next tick time

Schedule Status

  • ▶️ Running: Schedule is active
  • ⏸️ Stopped: Manually paused
  • 🔄 Refreshing: Updating schedule

Manual Schedule Control

  • Start/Stop: Toggle schedule on/off
  • Test: Run immediately regardless of schedule
  • Edit: Modify cron expression (requires code change)

Troubleshooting

Common Issues

"Asset not found" Error

  • Ensure all dependencies materialized
  • Check if asset definition exists
  • Reload definitions in UI

Meltano State Issues

  • DO NOT manually modify state
  • Contact support for state resets
  • Check logs for specific state errors

dbt Compilation Errors

  • Review dbt logs for model name
  • Check for missing dependencies
  • Validate warehouse permissions

Getting Run URLs

  1. Click on any run
  2. Copy URL from browser
  3. Share with support for help

Best Practices

Daily Operations

  1. Check Assets page for freshness
  2. Review any failed runs
  3. Monitor schedule execution
  4. Validate critical metrics

Development Workflow

  1. Test changes with single asset runs
  2. Validate in development branch first
  3. Monitor first production run closely
  4. Document any configuration changes

Next Steps