Confusing / missing stuff on Apache Airflow documenation
- worker
- Scheduler starts the executor on startup.
- Scheduler sends tasks to the executor.
- Default configured executor: runs a single task at a time.
- Runs out-of-the-box with local file-db.
- Parallelize task instances locally.
- Depends on real database backend like MySQL or Postgres.
- Scaling out of workers
- Depends on Celery
- Celery depends on a broker (RabbitMQ, Redis, AWS SQS)
- does celery need a db backend?