The distributed version of TimescaleDB, a time-series database on PostgreSQL, has been announced and is currently in private beta. The new architecture is designed to scale horizontally by chunking data across multiple nodes, rather than traditional sharding. This approach allows for greater flexibility and scalability, including the ability to add or remove nodes as needed, and to manage data retention policies, data tiering, and other features. Benchmarks have shown that the system can sustain high write rates of over 12 million metrics per second at 9 nodes. The distributed architecture is designed to provide a robust and familiar foundation for scale-out clustering, with a focus on usability and transparency. It leverages existing abstractions, such as hypertables and chunking, and provides elasticity, fault tolerance, and other operational features. While it does not overcome the CAP Theorem, which states that there is an implicit tradeoff between consistency, availability, and partition tolerance in distributed systems, TimescaleDB achieves high availability using replicated instances that perform prompt and automated recovery from failure.