A tool for benchmarking TimeScaleDB
Benchmarks takes the number of workers and a .csv file (.csv removed) with format "hostname, start_time, end_time" (header included) to generate and benchmark query run time. Queries are allotted to workers based on hostnames (no two workers share queries that touch the same hostnames).
Set up PostgreSql server.
go get github.com/mogarg/timescale
./benchmark -numWorkers 10 -file query_params- Testing
- Generating queries on the fly (currently uses a base query)
- Balanced Scheduling: Alot next hostname to some worker with the least amount of pending queries to run (could be useful for bigger workloads.
- More stats by percentiles.
- Any architectural changes suggested.
- Better errors handling if some queries fail.