Time-Series Database (TSDB)

A time-series database is built specifically for data points stamped with a time: metrics, sensor readings, prices, application events. Instead of storing rows the way a general-purpose database does, a TSDB exploits the fact that data arrives in time order and is rarely updated — so it compresses aggressively, expires old data cheaply, and answers 'average per minute over the last day' far faster than Postgres would. InfluxDB, TimescaleDB, Prometheus, and ClickHouse are common choices. The workload is append-heavy and query-by-range: you write huge volumes continuously and read recent windows or downsampled rollups, almost never a single random row. For SaaS builders, a TSDB is the backbone of product analytics, observability dashboards, and usage-based billing. Practical note: watch cardinality — the number of unique tag combinations (user × endpoint × region) can explode and wreck performance, so keep high-cardinality identifiers out of your metric labels.