[GH] (fluss/feature/fluss-perf): Workflow run "CI" failed!

[GitBox <gi...@apache.org>]

The GitHub Actions job "CI" on fluss.git/feature/fluss-perf has failed.
Run started by GitHub user platinumhamburg (triggered by platinumhamburg).

Head commit for run:
7aab65859c90cbbe5aaba39263adb3cb3cb1399f / 白鵺 <oc...@alibaba-inc.com>
Add fluss-microbench module with YAML-driven perf framework and JMH benchmarks

Unified benchmarking module combining YAML-driven performance testing
(write/lookup/scan/mixed workloads with resource sampling and reporting)
and JMH micro-benchmarks (Arrow channel, log scanner).

Key features:
- MICROBENCH_ROOT-based output layout (.microbench/runs/datasets/baselines)
- Auto-managed latest/previous symlinks per scenario for iterative comparison
- --diff-previous and --diff-baseline flags for regression detection
- Forked server log capture (logs/server-0.log) and run.log tee
- 5 built-in presets: log-append, kv-upsert-get, kv-agg-mixed/listagg/rbm32
- Contributor guide at website/community/dev/benchmarking.md

CLI design: workload parameters (records, threads, warmup, duration) are
YAML-only and cannot be overridden via CLI, ensuring every benchmark report
is statistically reliable. Warmup semantics are per-phase global (divided
by threads internally). Validator enforces warmup < records/2, effective
records per thread >= 10,000, and unique phase names per scenario.

Phase names support type-prefix convention (e.g. write-bulk, write-sustained)
resolved via longest-prefix matching to PhaseType. Preset tuning: thread
counts capped at 8/4, warmup 500K global for JIT C2 coverage, duration-based
write phases (5m) for stable measurement windows, scan warmup for cold-start
elimination.

Report URL: https://github.com/apache/fluss/actions/runs/24436034434

With regards,
GitHub Actions via GitBox