This screenshot tells everything. You don't have enough memory in your cluster and hence it spills to disk (memory swap utilization is high). You have really small cluster and your workload needs more memory. Just try to use cluster with more memory and your problem will be solved 🙂