Whats the k8's limitation here? You can do it however you like. e.g. make a fluentd daemonset, deploy on both nodes have them slurp up to s3. There are tons of

Whats the k8's limitation here? You can do it however you like. e.g. make a fluentd daemonset, deploy on both nodes have them slurp up to s3. There are tons of ways to handle the logs and k8s can either help or stay out of the way. https://github.com/fabric8io/fluent-plugin-kubernetes_metadata_filter https://github.com/fabric8io/fluent-plugin-kubernetes_metada...