Skip to main content

Deployment for Apache Airflows

Airflow deployment and administration covers running Airflow reliably: its components (scheduler, webserver or API server, metadata database, executor and workers), choosing an executor such as Local, Celery or Kubernetes, managing users and configuration, and monitoring the system once real pipelines depend on it.

Stage: Production · Part of the Apache Airflow roadmap · Last reviewed 19 November 2026

What to learn in Deployment

Deployment interview questions

What does the Airflow scheduler do?

The scheduler reads DAG files, creates DAG runs according to schedules, decides which task instances are ready based on dependencies and trigger rules, and sends them to the executor. It also handles retries. If the scheduler is down, no new runs or tasks start.

When would you choose the Kubernetes executor over Celery?

KubernetesExecutor starts each task in its own pod, giving isolation and per-task resource settings with no idle workers, which suits varied or bursty workloads. CeleryExecutor keeps a pool of long-running workers with lower task startup latency but needs a broker such as Redis or RabbitMQ.

Where Deployment fits in the Apache Airflow roadmap

Deployment belongs to the Production stage (advanced level). Operate.

Back to the full Apache Airflow roadmap