Elastic's new frozen data tier in Elasticsearch decouples compute from storage, utilizing low-cost object stores like Google Cloud Storage, Azure Blob Storage, or Amazon S3 to efficiently manage and query large volumes of data without rehydration. This innovative approach allows for scalable data management, enabling searches over vast datasets, such as petabyte-scale data, with significant cost savings. The frozen tier is designed for infrequent access scenarios, balancing performance with storage costs by leveraging an on-disk least-frequently-used (LFU) cache to optimize query speeds. Benchmark tests demonstrate that although the frozen tier's initial query performance is slower than traditional tiers, repeated queries benefit significantly from caching, achieving performance levels similar to those of the hot, warm, and cold tiers. This makes the frozen tier ideal for operational, security, and historical analyses, where fast access to large data sets is necessary without the overhead of maintaining vast amounts of local storage. The feature is integrated into Elasticsearch's index lifecycle management and is available for both self-hosted deployments and Elastic Cloud.