An kubernetes cluster is required to develop for Airflow if you intend to use kubernetesPodOperators. Minikube is a good option for this.
using helm, it's easy to install Airflow locally.
helm repo add apache-airflow https://airflow.apache.org
helm upgrade --install airflow apache-airflow/airflow --namespace airflow --create-namespace
to delete it from your k8s cluster:
helm delete airflow --namespace airflow
you'll need to tunnel into your running pods to view the web service:
kubectl port-forward svc/airflow-webserver 8080:8080 --namespace airflow
Then go to localhost:8080 in your browser.
to add a locally developed dag to your airflow instance, you can use kubectl to copy it in.
kubectl cp eddy-process-dag.py airflow-scheduler-66867df4fc-2jrtn:/opt/airflow/dags/. -n airflow
kubectl cp eddy-process-dag.py airflow-worker-0:/opt/airflow/dags/. -n airflow
Note you must copy it to both the scheduler AND the worker