Client hangs if cluster is closed/shutdown

I find that there is no way to query if the cluster is still live, and any tasks submitted will hang with no method for checking connection status.

Is there any workaround or heartbeat that occurs to prevent long hangs?