TimescaleDB

TimescaleDB is a Postgres extension that makes Postgres fast at time-series workloads: server metrics, IoT sensor readings, financial ticks, event logs. Not a separate database. An extension on top of Postgres. Your existing tools, ORMs, and knowledge all still work. C. The key feature: hypertables automatically partition your data by time, so queries like "average CPU usage over the last hour" are fast even on billions of rows. Continuous aggregates pre-compute rollups. Compression reduces storage by 90%+ for time-series data.

The community edition is free and open source for self-hosting. Installs as a Postgres extension. If you can install extensions, you can run TimescaleDB. Timescale Cloud (managed Postgres with TimescaleDB) starts at $0.023/hr (~$17/mo) for a small instance. Scales based on compute and storage. They've rebranded to Tigerdata recently.

Solo: self-host on your existing Postgres. Free. Small teams: self-host or Cloud at ~$17/mo to skip ops. Medium to large: evaluate Cloud vs self-hosted based on ops capacity. TimescaleDB is straightforward to self-host since it's just a Postgres extension.

The catch: some advanced features (multi-node, certain compression options) were previously under a commercial license; the licensing has shifted over the years. Check the current Timescale License for your use case. And if your time-series workload is pure metrics with Prometheus-style queries, VictoriaMetrics or InfluxDB might be more purpose-built.