Hello the airflow community ! We are heavily using airflow in my company as well as Kubernetes in Google Cloud and we got a really good experience with KubernetesPodOperator. It helps us to smoothly bring our machine learning pipelines from development to production. I just dropped out a medium article where I wanted to share that experience, and hopefully it might be useful to someone! So I believe it might be good idea to share it here as well: https://medium.com/dailymotion/bring-machine-learning-models-faster-to-production-with-airflow-and-kubernetes-e9d47ca3bee5
Your comments are most welcome ! Cheers, Eric
