SysPad › Lessons › database lessons › Tiger (TimescaleDB)

How Tiger (TimescaleDB) works

Tiger / TimescaleDB - a PostgreSQL time-series database (hypertables, columnar compression, continuous aggregates). Full Postgres SQL and relational joins, portable across clouds. Ingest-heavy; columnar compression speeds analytical scans. Sized by Tiger Cloud compute tier.

Time-series data, metrics, sensor readings, trades, piles up fast and crushes a normal table. You could learn a whole new specialised database, or you could keep Postgres. Tiger (TimescaleDB) is the second path: it turns PostgreSQL into a serious time-series engine, so you get time-series performance and the SQL you already know.

What you will learn

Why does ordinary time-series data overwhelm a normal table?
What is a hypertable, and how does chunking speed up queries?
How does columnar compression cut storage and accelerate analytics?
Why does "it is just Postgres" matter, and how does it differ from Amazon Timestream?

A hypertable: Postgres that partitions itself

Tiger (the database formerly called Timescale) adds a hypertable to Postgres. You read and write it like a single table, but it automatically partitions rows into chunks by time. A query for "last 24 hours" then touches only the relevant chunks instead of the whole history.

Columnar compression on history

Recent chunks stay row-based so inserts are fast. Older chunks get columnar compression, which shrinks storage dramatically and makes analytical scans (averages, sums over a column) much faster, since only that column is read.

⚓ IoT and sensor telemetry, application and infrastructure metrics, financial tick data, and any append-heavy timestamped workload.

Continuous aggregates stay fresh

Re-computing "hourly averages over a year" from raw rows every time is wasteful. Continuous aggregates maintain those rollups incrementally, updating as new data lands, so dashboards read a small, current summary instead of scanning raw history.

It is just Postgres (and portable)

Because Tiger is a Postgres extension, you keep full SQL, joins to your relational tables, and the entire Postgres ecosystem of drivers and tools. It runs as Tiger Cloud or anywhere Postgres runs, so it is portable across clouds.

Flashcards

What is a hypertable?: A Postgres table that auto-partitions rows into time-based chunks, so time-range queries prune to only the chunks they need.
What does columnar compression on older chunks buy you?: Much smaller storage and faster analytical scans, since only the queried column is read.
What problem do continuous aggregates solve?: They keep rollups incrementally up to date, so dashboards avoid re-scanning raw history for common summaries.
Why does "Tiger is just Postgres" matter?: Full SQL, joins, and the Postgres ecosystem, plus portability across clouds, no proprietary API.
Tiger vs Amazon Timestream?: Tiger = Postgres-based, portable, full SQL. Timestream = AWS-native, serverless, purpose-built, but not Postgres and not portable.

Sources

Open the SysPad canvas · Official Tiger (TimescaleDB) documentation