Time-Series Benchmarks
Benchmarks for infrastructure monitoring, IoT, and observability workloads - the workloads that time-series databases (TimescaleDB, InfluxDB, QuestDB) and columnar engines are routinely asked to handle.
Why Time-Series Benchmarks?
Time-series workloads have distinct shapes that generic OLAP benchmarks do not stress:
- Time-partitioned access patterns - queries are almost always bounded by a time window, which interacts with partitioning and retention strategies.
- High-cardinality tag sets - host / sensor / metric identifiers create secondary indexes and demand fast dimension lookups.
- Downsampling and rollups - minute-level data is often queried hourly or daily; engines differ sharply on materialised aggregate performance.
- Append-heavy ingest - writes dominate, with rare updates or deletes.
Time-Series Benchmarks in BenchBox
| Benchmark | Data Source | Focus |
|---|---|---|
| TSBS DevOps | Simulated infrastructure monitoring (TSBS) | Time-series aggregation, host tags, DevOps dashboards, 18 queries |
When to Use
- TSBS DevOps - evaluate time-series databases (TimescaleDB, InfluxDB, QuestDB) and columnar engines (DuckDB, ClickHouse) on DevOps monitoring workloads. The queries exercise single-host lookups, multi-host rollups, and full-fleet aggregations.
Included Benchmarks
See Also
- Real-World Data Benchmarks - Real-world datasets (NYC Taxi, Flight Data)
- Industry Benchmarks - Vendor / practitioner benchmarks