Apache Airflow on Railway
Python pipeline orchestration with scheduler, API server and Celery workers on Railway.
Apache Airflow is an open-source platform for authoring, scheduling and monitoring data pipelines as Python code. You define workflows as DAGs (directed acyclic graphs) of tasks, and Airflow runs them on a schedule or on events, with retries, run history and a web UI.
Data engineers use it to orchestrate ETL and ELT jobs, dbt and Spark runs, and machine-learning pipelines. It is often chosen over cron or ad-hoc scripts when jobs have dependencies and need to be observed and re-run. Prefect and Dagster cover similar ground.
Airflow is a project of the Apache Software Foundation and is licensed under Apache-2.0.
What Apache Airflow can do
- DAGs defined in Python, with scheduling, retries and SLAs
- Web UI for runs, logs and task state
- REST API and Task Execution API (Airflow 3)
- Celery executor for distributed workers
- Deferrable operators and sensors via the triggerer
- DAG bundles that load DAGs from a git repository
What people use it for
- Nightly and hourly ETL/ELT jobs into Postgres or a warehouse
- Orchestrating dbt, Spark or Python batch jobs
- Machine-learning feature, training and batch scoring pipelines
- Event-driven workflows with sensors
- Replacing scattered cron jobs with one observable scheduler
Videos
Running Apache Airflow on Railway
The Railway template deploys Airflow 3 as separate services: API server (the only public one), scheduler, DAG processor, triggerer and Celery workers, plus Postgres for metadata, Redis as the Celery broker and a Railway Bucket for task logs. A one-shot init service migrates the database and creates the admin account. DAGs are delivered from a git repository through a DAG bundle, because volumes cannot be shared between Railway services. Cost follows the CPU and memory of the several always-running services.
Apache Airflow templates on Railway
Tracklayer templates are listed first. Other Apache Airflow templates follow, ordered by total deployments on Railway. How this works.
Deployment numbers as of 2026-11-20.
Questions about Apache Airflow
How do I get my DAGs into Airflow on Railway?
The template reads DAGs from a git repository through an Airflow 3 DAG bundle. Set the repository URL variable to your own repo and push changes to deploy. If it is left empty, example DAGs baked into the image are used.
Why are there so many services?
Airflow 3 splits its roles: the API server serves the UI and API, the scheduler queues tasks, the DAG processor parses DAG files, the triggerer runs deferrable operators and Celery workers execute tasks. The template follows the topology the project recommends for production.
Where do task logs go?
To a Railway Bucket (S3-compatible), so logs survive redeploys and can be read from any service.
Is Airflow free for commercial use?
Yes. Apache Airflow is Apache-2.0 licensed. You pay only for the Railway resources the services use.
Guides
Similar apps
Tracklayer maintains the templates marked as such; Apache Airflow is a trademark of its owners and this site is not affiliated with them.