Ingestion related benchmarks ( tsbs_load*...) should expose the per command/operation latency summary

timescale / tsbs

Time Series Benchmark Suite, a tool for comparing and evaluating databases for time series data

MIT License

1.26k stars 297 forks source link

Ingestion related benchmarks ( tsbs_load*...) should expose the per command/operation latency summary #159

Open filipecosta90 opened 3 years ago

filipecosta90 commented 3 years ago

Currently, on the ingestion benchmarks, we're only able to assess and compare the different solutions based upon the ingestion rate and total ingestion time. As important as the command rate is the command latency ( specifically on writes for TSDBs ). Given this is expected to be the de-facto standard for TSDBs benchmarks I believe we're missing that major feature. PS: If you guys agree I'm completely up for pushing a PR that does exactly that.

filipecosta90 commented 3 years ago

@jonatas will hapilly submit a PR over this feature. I believe we're missing critical info on ingestion ( which is a big part on TSDBs worflows ).

jonatas commented 3 years ago

I totally agree @filipecosta90! The PR is very welcome :)