TimescaleDB Documentation (Tiger Data)

What TimescaleDB is

TimescaleDB is a time-series database packaged as a PostgreSQL extension. Time-series data is a stream of measurements with a timestamp: temperature every minute, a stock price every second, API latency per request. You keep writing ordinary SQL against ordinary Postgres, and TimescaleDB adds features for this kind of data. The company behind it renamed itself from Timescale to Tiger Data in 2025; the extension is still called TimescaleDB.

What the docs cover

• Hypertables: tables that look normal but are split into time-based chunks automatically, so recent data stays fast to insert and query.
• Time buckets: time_bucket() groups readings into 5-minute, hourly or daily buckets for charts.
• Continuous aggregates: materialised views that update incrementally, so a dashboard of hourly averages does not rescan raw data.
• Compression and the columnstore: storing older data in a compressed, column-oriented form that is much smaller and fast to scan.
• Retention policies: dropping raw data older than, say, 90 days while keeping the daily summaries.
• Installation: self-hosting the extension on your own Postgres, or using Tiger Data's managed cloud.

Is it worth learning?

If your project collects readings over time (IoT sensors, app metrics, prices, energy meters), yes. It is the gentlest path into time-series work because everything else stays PostgreSQL: joins with your users table, backups, drivers and ORMs. Dedicated engines such as InfluxDB are the alternative when you want a separate system built only for metrics.

When you do not need it

A few thousand readings a day fit happily in a plain Postgres table with an index on (device_id, recorded_at). Add TimescaleDB when the table grows into the tens of millions of rows or the dashboard queries start to slow down.

Prerequisites and time

Comfortable SQL and a local PostgreSQL install. A weekend to load a year of sample readings, build hourly aggregates and set up compression.

Free or paid?

The extension is free to self-host; some advanced features have their own licence terms, so check the docs for the feature you need. The managed cloud is paid, with a trial.

Best for: developers building IoT, monitoring or finance features who already know PostgreSQL.