Concepts & ArchitectureLesson 04

Airflow Web UI vs. CLI

Web interface.

Access via browser: http://localhost:8080

Essential CLI Commands

Docker Access

hljs bash
# Enter the Airflow container
docker compose exec -it airflow-webserver bash

Installing Python Packages

Airflow tasks run inside Airflow containers, so any extra Python library must be available inside those containers before the DAG runs.

For a quick local test, open the Airflow webserver container and install the package manually:

hljs bash
docker compose exec airflow-webserver bash
pip install pandas

For a repeatable Docker setup, place the package requirement in your .env file or in the docker-compose.yml environment section, then restart the Airflow containers.

hljs bash
# .env file style
_PIP_ADDITIONAL_REQUIREMENTS=pandas

# docker-compose.yml environment style
_PIP_ADDITIONAL_REQUIREMENTS: "pandas"

System Diagnosis

hljs bash
# Show environment, DB, and executor info
airflow info

# Show help menu
airflow -h

# Check DAGs folder location
cat airflow.cfg | grep "dags_folder = "

Executor

hljs bash

cat airflow.cfg | grep "executor = "

airflow info | grep executor

DAG Management

hljs bash
# Check Python syntax errors in DAG file
python3 my_dag.py

# List all loaded DAGs
airflow dags list

# List DAGs that failed to load (syntax/import errors)
airflow dags list-import-errors

Testing & Execution

hljs bash
# Test a specific task WITHOUT saving to DB
airflow tasks test <dag_id> <task_id> <date>

# Example:
airflow tasks test crypto_pipeline fetch_prices 2026-04-20


# Manually trigger a full DAG run
airflow dags trigger -e <date> <dag_id>