Who is responsible for creating DAGs and assigning tasks in Airflow, and is this part of the runtime?

In Apache Airflow, the DAG (Directed Acyclic Graph) is defined by a developer or data engineer in a Python file. The DAG file declares the workflow structure, including the tasks and their dependencies. The tasks themselves are also defined within the DAG file, typically as instances of operators. The Airflow scheduler, which runs as a separate process, reads these DAG files, parses them, and schedules the tasks for execution. The actual execution of tasks happens in worker processes or executors, which are part of the runtime environment. Therefore, the creation of DAGs and tasks is a design-time activity performed by developers, while the scheduler and executor handle the runtime execution of those tasks.