DAG deployments
We’ll start with a discussion of how to get your Airflow DAGs to your Airflow deployment. We won’t discuss specific implementation details because they can vary depending on your specific use case, security context, and operational needs.
Bundling
The DAG bundling pattern may be the most obvious pattern and is a very common starting point on your journey to using Airflow in production. It’s also antithetical to the code as configuration paradigm that is core to Airflow’s development and consumption.
The general pattern is to physically bundle your DAGs within the same file system as your Airflow system, even in a system with distributed compute. This is achieved by having your container build process install Airflow and copy your DAGs into the container during the image build process. This image can then be tagged and distributed to ensure that all components are operating using the exact same Airflow, plugin, and DAG versions.
...
