TimescaleDB is an extension that auto-partitions a table by a time column to give you faster queries based on time ranges. This is great for timeseries data (which is just data with time as part of the primary key) and allows for fast ingest with recent time partitions being smaller and in-memory.

PipelineDB is a Postgres build (and soon to be extension) that runs real-time continuous "queries" on incoming data as materialized views. This is similar to the Kafka KSQL or Spark streaming and is effectively taking in a constant stream of data and re-running a query incrementally to give you up-to-date analytics with windowing.

Neither products are full-scale data warehouse solutions. TimescaleDB is currently limited to time-based partitioning, single-node only, and still uses standard row-based tables. PipelineDB discards raw incoming data and only keeps materialized views so you must know your queries beforehand. Both give you the advantage of using full SQL and the rest of the postgres ecosystem to easily join and analyze with other operational data without doing lots of ETL.

Another worthy mention is CitusDB which is another auto-sharding extension to Postgres, mainly focused on scaling horizontally for OLTP scenarios.