Strange thing is I see these netty disconnects only on Master and Coordinating nodes. I dont see any issues on Data nodes as such. They are not crashing and are able to join back the Cluster quickly. The Cluster was up and running and working well for almost a year now and didnt see any such issues as such.