How to use the DBT 3 test kit
summary
DBT3 is a TPCH-derived workload that generates data from gigabytes to tens of terabytes and runs three stages: parallel load, a 21-query power test, and multi-stream throughput tests (configurable streams and RNG seeds). Output includes explain/analyze plans, per-query profiles, and extensive OS/DB counters — collects SAR and pidstat and charts CPU, IO, buffer hits, table/index stats. Examples include measuring BRIN vs B-tree maintenance overhead and a scalability gain from ~12–14 to ~20 concurrent streams; installation and scripting are manual and brittle.
AI-written summary · about
chapters (11)
- 0:00Start
- 3:57Decision Support Benchmark
- 5:27Power Test
- 6:26Throughput Test
- 8:25Tpc Benchmarks
- 25:41Load Time
- 26:24Query Processing Power
- 27:31Composite Score
- 35:03Query Plans
- 36:13Plan Disaster
- 38:20Power and the Throughput Test
related talks
49mdbt: A Software Engineering Approach To The Analytics Workflow
51mIntroduction to Fair-Use TPC Benchmarking Kits
44mDemystifying Benchmarks: How to Use Them To Better Evaluate Databases
28mdbt-score: a linter for your dbt model metadata
31mDemanding the impossible rigorous database benchm
01h 07mHow to Count Things with dbt - Data Build Tool
02mDBTB INT Thomas Trolez r
48mDBtune: AI-driven performance tuning for all PostgreSQL flavors