This is an automated email from the ASF dual-hosted git repository.
nzw921rx pushed a commit to branch dev
in repository https://gitbox.apache.org/repos/asf/seatunnel.git
The following commit(s) were added to refs/heads/dev by this push:
new 66c9b4b762d [Fix][CI] Default benchmark diagnostics to Java 11 (#13259)
66c9b4b762d is described below
commit 66c9b4b762d645ba34947336f7a6b192501a7ab7
Author: zhiweiniu <[email protected]>
AuthorDate: Mon Nov 16 14:09:33 2026 +0800
[Fix][CI] Default benchmark diagnostics to Java 11 (#13259)
---
.github/workflows/benchmarks_diagnostics.yml | 4 ++--
docs/en/engines/zeta/benchmark.md | 2 +-
docs/zh/engines/zeta/benchmark.md | 2 +-
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/.github/workflows/benchmarks_diagnostics.yml
b/.github/workflows/benchmarks_diagnostics.yml
index 8b10adb5ce4..5d43576520d 100644
--- a/.github/workflows/benchmarks_diagnostics.yml
+++ b/.github/workflows/benchmarks_diagnostics.yml
@@ -39,9 +39,9 @@ on:
required: true
type: choice
options:
- - '8'
- '11'
- default: '8'
+ - '17'
+ default: '11'
profile:
description: 'Profiler mode. all runs CPU, wall-clock, lock, and GC
separately.'
required: true
diff --git a/docs/en/engines/zeta/benchmark.md
b/docs/en/engines/zeta/benchmark.md
index f67efac238e..c23e990ec79 100644
--- a/docs/en/engines/zeta/benchmark.md
+++ b/docs/en/engines/zeta/benchmark.md
@@ -286,7 +286,7 @@ selectors such as `.*` or a class name that matches several
methods are rejected
| `seatunnel_ref` | Branch, tag, or SHA to diagnose when no PR is selected. |
| `pr_number` | Optional trusted PR ID. When set, its head replaces
`seatunnel_ref` as the diagnostic target. |
| `benchmark` | One method selector, `<benchmark-method>$`, obtained from
`-l`. Selectors matching multiple methods are rejected. |
-| `java_version` | `8` or `11`; use the JDK of the normal run being
investigated. |
+| `java_version` | `11` or `17` (default `11`); use the JDK of the normal
benchmark run being investigated. |
| `profile` | `cpu` for execution hot spots, `wall` for elapsed-time stacks
including waits, `lock` for lock contention, `gc` for allocation/GC metrics, or
`all` for separate runs of all four. |
| `capture_jfr` | Enable to add a separate JFR recording for offline analysis.
|
| `jmh_args` | Optional JMH arguments or workload parameters, using the
selected method's `-lp` output. Leave empty for defaults; forks remain fixed at
1. |
diff --git a/docs/zh/engines/zeta/benchmark.md
b/docs/zh/engines/zeta/benchmark.md
index db03be067b2..505fca4ee71 100644
--- a/docs/zh/engines/zeta/benchmark.md
+++ b/docs/zh/engines/zeta/benchmark.md
@@ -268,7 +268,7 @@ Java 11 和 Java 17 分别执行这组对比,报告汇总每个版本的两轮
| `seatunnel_ref` | 未选择 PR 时,要诊断的分支、Tag 或 SHA。 |
| `pr_number` | 可选可信 PR 编号;填写后以 PR Head 替代 `seatunnel_ref` 作为诊断目标。 |
| `benchmark` | 填写从 `-l` 查询到的方法选择器 `<benchmark-method>$`;匹配多个方法的选择器会被拒绝。 |
-| `java_version` | `8` 或 `11`,与待分析的正常运行保持一致。 |
+| `java_version` | `11` 或 `17`(默认 `11`);与待分析的正常基准测试运行使用相同的 JDK。 |
| `profile` | `cpu` 查看执行热点,`wall` 查看含等待在内的耗时栈,`lock` 查看锁竞争,`gc` 查看分配与 GC
指标;`all` 分别运行这四种模式。 |
| `capture_jfr` | 勾选后增加一次独立 JFR 录制,用于离线分析。 |
| `jmh_args` | 可选 JMH 参数或负载参数,以所选方法的 `-lp` 输出为准;留空使用默认值,fork 固定为 1。 |