TimescaleDB Time-Series Database

TimescaleDB is a free and open-source time-series database optimized for fast ingest and complex queries, built as an extension on PostgreSQL. Created by Timescale Inc. (founded by Michael J. Freedman, Ajay Kulkarni, and Mike Freedman in 2017, headquartered in New York City and Boston, backed by over $180 million in total funding including a $40M Series C in 2020 led by Tiger Global), TimescaleDB combines the scalability of purpose-built time-series databases with the full SQL power and reliability of PostgreSQL. Key features: hypertables: the core abstraction. A hypertable appears as a single PostgreSQL table but is automatically partitioned into chunks across time intervals and (optionally) space. Queries work transparently, but the engine routes them to only relevant chunks, dramatically improving performance for time-range queries. Automatic partitioning: TimescaleDB automatically creates new chunks as data arrives. Each chunk corresponds to a specific time interval and is implemented as a regular PostgreSQL table. Chunk size is optimized for in-memory processing. Compression: columnar storage with type-specific encoding (delta-delta encoding for integers, Gorilla encoding for floats) reduces storage costs by up to 20x. Compressed chunks can still be queried transparently. Continuous aggregates: materialized views that are automatically maintained in the background as new data arrives, enabling fast queries on rollups without re-scanning raw data. Data retention policies: automatically drop old data after a configurable time period. PostgreSQL compatibility: since TimescaleDB is a PostgreSQL extension, it supports all PostgreSQL features including JOINs, transactions, triggers, foreign keys, full-text search, JSONB, PostGIS, and all client libraries. Functions: time_bucket (GROUP BY time intervals), gapfilling, first/last, and time-series-specific aggregates. C. Apache-2.0/TSLA.