Weaviate 1.36 Release
Blog post from Weaviate
Weaviate v1.36 introduces significant updates, including the new HFresh vector index type, which is designed to improve memory efficiency and write throughput for large datasets by storing vectors on disk rather than in memory. Several features have moved to general availability, such as server-side batching, which optimizes data flow without manual configuration, and Object Time-to-Live, which automates data expiration to support lifecycle management. Enhancements to asynchronous replication provide more control over data propagation, while the ability to drop inverted indices helps manage disk space efficiently. Backup restoration cancellation is now available, allowing users to abort ongoing restore operations to free resources. These updates collectively enhance Weaviate's indexing performance, data management capabilities, and operational reliability at scale, and the release is available both open-source and on Weaviate Cloud.