Query performance benchmark #16
Description
Is your feature request related to a problem? Please describe
In previous internal benchmarks, query performance was roughly compared by manually ingesting data for several days and navigating different pages of the Jaeger Frontend to subjectively experience the query duration differences. Although VictoriaTraces performed comparably to ClickHouse and significantly outperformed Elasticsearch, these results were neither accurate nor quantifiable.
Describe the solution you'd like
It's good to design re-runnable benchmark to measure the query duration for different scenarios:
- Filter traces by condition.
- Filter trace by
trace_id.
Given that the time range filter significantly impacts performance, the benchmark should target queries across large time ranges, short time ranges, fresh data, old data, and more.
Describe alternatives you've considered
No response
Additional information
No response