Ah, I see. Yes, that should definitely be good enough (at least for a first
pass).

Do you know how most users tend to deploy and use Flink in practice?
Job-scoped clusters seem preferable for many reasons, but it's possible
that that's only really practical when using YARN/Kubernetes/Mesos. Would
we be be cutting out a large population by only "supporting" single-job
clusters?

On Sun, Dec 10, 2017 at 3:42 PM, Aljoscha Krettek <[EMAIL PROTECTED]>
wrote:
--
-Ben