that's not an uncommon scenario. What worked for me so far is to encode the cluster identifier in the index names (such as logs-cluster-${cluster_id}) as well as add a field upon ingestion that contains the cluster identifier. If you're using Filebeat or Elastic Agent you could achieve the latter using the add_kubernetes_metadata processor, which can add labels of the monitored resource to each document.
This then provides the flexibility to limit the queried indices to a cluster or query all of logs-* but filter or aggregate over the cluster identifier in the labels.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.