[Java] [Benchmarking] Large stdout when running benchmarks
- Dominant language
- Java
- Stars
- 94
- Forks
- 152
- Avg merge
- 3d 16h
- Merged PRs (30d)
- 11
Description
Since ARROW-15058 running the Java benchmarks results in stdout logs that are extremely large (we're seeing them at ~14gb). This is disrupting running benchmarks on conbench.
We've temporarily worked around this by disabling stdout logging entirely, but this will make debugging issues in the future much harder.
Is there any way to reduce the volume of stdout logging when running these benchmarks?
cc @davisusanibar [~fan_li_ya]
**Reporter**: [Elena Henderson](https://issues.apache.org/jira/browse/ARROW-15259) / @ElenaHenderson
**Note**: *This issue was originally created as [ARROW-15259](https://issues.apache.org/jira/browse/ARROW-15259). Please see the [migration documentation](https://github.com/apache/arrow/issues/14542) for further details.*
Contributor guide
Research direction
The issue does not name a source file, benchmark entry point, or test. Start by reproducing a Java benchmark run and tracing the logging introduced since ARROW-15058; compare output volume with the temporary stdout-disabled workaround. Done means substantially smaller benchmark stdout while retaining useful debugging output and avoiding disruption to conbench.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- java
- Domain
- performance
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100