dagster This framework has a lot in common with Apache Airflow.
The scheduler and UI are divided into different modules. Work with s3 resources and local files is available.
Dagster implements a concept with work, creation and cleaning upon completion of work. There is also a quick scheduler here.
Argo Workflows This solution works on Go. Containers are launched in Kubernetes.
It is convenient to use because of the isolation of virtual environments. However, there is a difficulty in implementing testing.
It is necessary to run pipelines on Go, in which datasets in python will be compared.