We are facing some issues with our spark application that loads data from s3 to elasticsearch. Since some weeks these jobs are not closing off correctly anymore in Kubernetes.