With our custom columnar format and indexes, we are able to filter relevant data files where high cardinality column is present. This helps us to keep the queries faster for high cardinality labels as well, thus, allowing us to quickly drill down on specific pod_id/cluster_id/customer_id kind of labels.