I think the answer here is maybe. See
this. I am thinking that you can use a resource manager to send that work to another cluster, but another user reported to me it did not work. I suggest you try, and if it fails, just use a flow of flows with each subflow on a different executor