HI I am facing similar error, the cluster runs 24/7 and this issue is observed for a few runs in the day. The data volume being processed is not huge but the logic that this needs to go though is complex. I do not want to include the display in production code. Please let me know your thoughts