Features

Dagu provides a comprehensive set of features for building and managing workflows. This section covers everything you need to know about each feature in detail.

Feature Categories

🖥️ Interfaces

How to interact with Dagu:

• Command Line Interface - Run and manage workflows from the terminal
• Web UI - Monitor and control workflows visually
• REST API - Integrate Dagu programmatically

🔧 Executors

Different ways to run your commands:

• Shell - Run any command or script (default)
• Docker - Execute in containers for isolation
• SSH - Run commands on remote servers
• HTTP - Make REST API calls and webhook requests
• Mail - Send email notifications and reports
• JQ - Process and transform JSON data

⏰ Scheduling

Control when workflows run:

• Cron expressions with timezone support
• Multiple schedules per workflow
• Start/stop/restart patterns
• Skip redundant executions

🚀 Execution Control

Advanced execution patterns:

• Parallel execution with concurrency limits
• Conditional execution (preconditions)
• Continue on failure patterns
• Retry and repeat policies
• Output size management
• Signal handling

📊 Data Flow

Managing data in workflows:

• Parameters and runtime values
• Output variables between steps
• Environment variable management
• JSON path references
• Template rendering
• Special system variables

📋 Queue System

Workflow orchestration at scale:

• Built-in queue management
• Per-DAG queue assignment
• Priority-based execution
• Manual queue operations
• Concurrency control

📧 Notifications

Stay informed about workflow status:

• Email alerts on success/failure
• Custom notification handlers
• Log attachments
• Flexible SMTP configuration

🔍 OpenTelemetry Tracing

Distributed tracing for observability:

• End-to-end workflow visibility
• Performance bottleneck identification
• Nested DAG correlation
• Integration with Jaeger, Tempo, etc.

Distributed Execution

Scale workflows across multiple machines:

• Coordinator-worker architecture
• Label-based task routing
• Real-time worker monitoring
• Requires shared storage for DAG files and state
• Horizontal scaling

Worker Labels

Task routing for distributed execution:

• Capability-based worker tagging
• Flexible label matching
• Resource optimization
• Geographic distribution
• Compliance requirements

Feature Highlights

Zero Dependencies

Unlike other workflow engines, Dagu requires:

• No database
• No message broker
• No runtime dependencies
• Just a single binary

Language Agnostic

Run anything that works on your system:

steps:
  - name: python
    command: python script.py

  - name: node
    command: npm run task

  - name: go
    command: go run main.go

  - name: bash
    command: ./script.sh

Hierarchical Workflows

Build complex systems from simple parts:

steps:
  - name: data-pipeline
    run: etl
    params: "DATE=today"
    
  - name: ml-training
    run: train
    parallel: "image text"
    params: "MODEL=latest TYPE=${ITEM}"
    
  - name: deployment
    run: deploy
    parallel: "staging production"
    params: "ENV=${ITEM}"

---
name: etl
params:
  - DATE: today
steps:
  - name: etl
    command: python etl.py

---
name: train
params:
  - MODEL: latest
  - TYPE: ""
steps:
  - name: train
    command: python train.py --model ${MODEL} --type ${TYPE}

---
name: deploy
params:
  - ENV: ""
steps:
  - name: deploy
    command: kubectl apply -f deployment.yaml --env ${ENV}

See Also

Explore specific features:

• Command Line Interface - Master the CLI
• Shell Executor - Run commands effectively
• Scheduling - Automate workflow execution
• Execution Control - Advanced patterns