TimescaleDB 2.5 is a multi-node, petabyte-scale relational database for time-series data that offers improved performance and scalability. It supports distributed hypertables, which enable users to run large-volume time-series workloads across multiple physical instances. The new release introduces continuous aggregates on distributed hypertables, allowing users to compute aggregates in real-time while maintaining the illusion of a single continuous table across all time. This feature enables fast query performance by pre-computing results and storing them on the access node, reducing the need for raw data retrieval at query time. Additionally, TimescaleDB 2.5 includes support for PostgreSQL 14, new timezone management capabilities, and various bug fixes.