Dataflow is designed to work well with kubectl, rather than needing its own CLI (though maybe we'll add one
someday 😀).
Task you can do with kubectl:
Create a pipeline:
kubectl apply -f examples/101-hello-pipeline.yaml
Wait for the pipeline to be running:
kubectl wait pipeline/101-hello --for=condition=running
Restart pipeline:
kubectl delete pod -l dataflow.argoproj.io/pipeline-name=xxx