TimescaleDB

When to Use

• SQL familiarity: You want time-series but already know SQL and use Postgres drivers.
• Relational + Time: You need to JOIN your sensor data (Time Series) with Device metadata (Relational Tables).
• Compression: Highest-in-class compression (90%+) for historical data.

Quick Start

-- Convert standard table to hypertable
SELECT create_hypertable('conditions', 'time');

-- Query using standard SQL time-bucket functions
SELECT time_bucket('15 minutes', time) AS bucket,
       avg(temperature)
FROM conditions
GROUP BY bucket
ORDER BY bucket DESC;

Core Concepts

Hypertables

Continuous Aggregates

Compression

Best Practices (2025)

• Enable Compression: It improves query speed (less I/O) and saves massive disk space.
• Use Tiered Storage: Keep recent hot data on SSD, move compressed old data to S3 (Bottomless storage in cloud).
• Join tables: Use the power of Postgres to join your metrics with your business data.

• Don't update compressed chunks: Updating old, compressed data is slow (Copy-on-write). Design for append-only patterns.

References

• TimescaleDB Documentation