This is an automated email from the ASF dual-hosted git repository.
Aias00 pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/hertzbeat.git
The following commit(s) were added to refs/heads/master by this push:
new 7c5c10cdde [doc] add new committer blog for orangeCatDeveloper (#4407)
7c5c10cdde is described below
commit 7c5c10cdde7c7c5d053b59ac426ef1b093103e1c
Author: NekoPunch <[email protected]>
AuthorDate: Sun Sep 27 18:02:02 2026 -0700
[doc] add new committer blog for orangeCatDeveloper (#4407)
---
home/blog/2026-09-23-new-committer.md | 78 ++++++++++++++++++++++
home/docs/help/alert_sms.md | 2 +-
.../2026-09-23-new-committer.md | 57 ++++++++++++++++
.../current/help/alert_sms.md | 2 +-
home/src/pages/team/data/member.json | 5 ++
5 files changed, 142 insertions(+), 2 deletions(-)
diff --git a/home/blog/2026-09-23-new-committer.md
b/home/blog/2026-09-23-new-committer.md
new file mode 100644
index 0000000000..4ca10dac57
--- /dev/null
+++ b/home/blog/2026-09-23-new-committer.md
@@ -0,0 +1,78 @@
+---
+title: Welcome HertzBeat's New Community Committer!
+description: orangeCatDeveloper describes starting from long-standing issues
to fix data reliability, security, and deployment problems and add
infrastructure monitoring on the way to becoming a Committer.
+author: orangeCatDeveloper
+author_title: orangeCatDeveloper
+author_url: https://github.com/orangeCatDeveloper
+tags: [community]
+keywords:
+ [
+ open source monitoring system,
+ alerting system,
+ AI observability,
+ Apache,
+ Apache Committer,
+ HertzBeat,
+ ]
+cover_headline: Welcome orangeCatDeveloper
+cover_kicker: New Committer
+---
+
+> Hello everyone, it's a great honor to have been elected an Apache HertzBeat™
Committer by a vote of the project's PMC.
+
+## Self-Introduction
+
+I'm [orangeCatDeveloper](https://github.com/orangeCatDeveloper), a software
engineer focused on AI infrastructure.
+What I care about is the less visible half of AI systems: the data pipelines
that feed models, the execution environments that let AI agents run safely, and
the observability that tells you whether all of it is working.
+
+## First Encounter with Apache HertzBeat
+
+What first made me look closely at HertzBeat was its AI side. It ships with AI
chat and an MCP Server, so a large language model can query monitoring data,
create monitors, and analyze alerts on its own.
+HertzBeat is a top-level project of the Apache Software Foundation. As of
September 2026, it has more than 7,400 GitHub stars, over 300 contributors, and
more than 220,000 pulls of its official Docker image.
+Every bug I fix in the collector makes its way into monitoring systems around
the world with the next release.
+
+Once I started reading the code and the issues, though, I ended up looking at
the other end of the pipeline. However smart the model is, all it ever sees are
the numbers the collector hands over.
+If those numbers are wrong, it will confidently reach the wrong conclusion.
The issue list had quite a few problems like that, some open for two or three
years, so that is where I started.
+
+## The Path to Open Source Contributions
+
+As of September 2026, I have [37 merged
PRs](https://github.com/search?q=is%3Apr+is%3Amerged+author%3AorangeCatDeveloper+repo%3Aapache%2Fhertzbeat+repo%3Aapache%2Fhertzbeat-helm-chart&type=pullrequests)
in HertzBeat. They closed 28 issues, 19 of which had been open for more than a
year; the oldest dated back to January 2023.
+A collector is a data pipeline in the plainest sense: it pulls numbers from
more than a hundred kinds of systems and hands them to alerts, dashboards, and
now LLMs.
+Most of my work has been making that pipeline trustworthy.
+
+The clearest example was a problem first reported in 2023: restart HertzBeat
and every website monitor reports slow responses, while the websites themselves
are fine.
+The root cause was that after a restart every collection job fired in the same
second and fought over resources.
+After reproducing it, I added a random jitter to each job's first run so they
no longer fire all at once, and under the same test load the p50 response time
the monitors reported dropped from 2865ms to 617ms and p95 from 6801ms to
1497ms.
+Problems like this never raise an error; they just quietly distort the data,
and alerts, dashboards, and AI analysis all read that data.
+Similar fixes covered a monitor's history splitting into a new series on every
edit, and edge-case data being silently dropped.
+
+Security is the other side. A monitoring system usually sits deep inside the
network, holding credentials for many servers and databases, so it must not be
the weakest link.
+The project depended on Bouncy Castle 1.68 from an abandoned release line,
which carried seven published CVEs that would never be patched. I migrated it
to the maintained 1.85, which has no known vulnerabilities.
+I also fixed certificate monitoring still checking the hostname with
verification turned off, and added a certificate verification toggle for email
servers, which helps with self-signed certificates on internal networks.
+
+Some of the problems were about deployment. At one point a database migration
collided with the automatic table creation at startup, so a fresh install got
stuck on boot and never came up again.
+I made the migration scripts safe to run repeatedly and let instances that
were already stuck recover on their own.
+In the Helm chart repo, I got CI running again and added startup probes for
Kubernetes deployments.
+I also added etcd monitoring; etcd is the core store of the Kubernetes control
plane.
+
+## Community Engagement and Growth
+
+The biggest thing I learned over this time is how to write a PR that saves the
reviewer effort: say what the user ran into, then why, then paste the real
output before and after the fix.
+Everyone's time is limited, and when the evidence is right there, the
discussion moves much faster.
+
+As a Committer, I want to spend more time on reviews and help newcomers get
their first PR merged.
+On my own side, there are three things I want to do next: keep clearing the
backlog of old issues;
+regularly audit dependencies that are abandoned or carry known
vulnerabilities, so supply chain risks are stopped before a release;
+and keep filling in monitoring coverage at the infrastructure layer, so the
data AI analysis relies on is both broader and more reliable.
+
+## Advice for Open Source Developers
+
+If you don't know where to start, try reading the issue list from oldest to
newest. Old issues are real problems users have hit, and many have been
discussed long enough to point toward a direction. Reproduce first; some will
turn out to be already fixed, and the ones that remain are worth the effort.
+Once you fix one, paste the before/after. It beats any amount of explanation.
+
+Use AI tools if they help, but make sure you can explain why every line you
submit is written the way it is.
+
+## Conclusion
+
+Thanks to everyone in the community for the patient reviews on every PR, and
to the PMC for the nomination and vote.
+HertzBeat is heading toward being an AI-driven observability platform, and I
hope every number it hands to the AI holds up to scrutiny. Let's keep going
together.
diff --git a/home/docs/help/alert_sms.md b/home/docs/help/alert_sms.md
index ed08670b38..c4894b840b 100644
--- a/home/docs/help/alert_sms.md
+++ b/home/docs/help/alert_sms.md
@@ -106,7 +106,7 @@ Now you can configure this information in your hertzbeat
application.
### Smslocal SMS Configuration
-SMSLocal is an all-in-one SMS service for businesses, with features like
multi-way sending, strong security, and 24/7 support. You can refer to
smslocal's [Developer Documentation](https://www.smslocal.com/developer/) for
configuration.
+SMSLocal is an all-in-one SMS service for businesses, with features like
multi-way sending, strong security, and 24/7 support. You can refer to
smslocal's [Developer Documentation](https://www.smslocal.com/resources/docs/)
for configuration.
Add/Fill in the following Smslocal configuration to `application.yml` (replace
parameters with your own SMS server configuration):
diff --git
a/home/i18n/zh-cn/docusaurus-plugin-content-blog/2026-09-23-new-committer.md
b/home/i18n/zh-cn/docusaurus-plugin-content-blog/2026-09-23-new-committer.md
new file mode 100644
index 0000000000..74b91244ad
--- /dev/null
+++ b/home/i18n/zh-cn/docusaurus-plugin-content-blog/2026-09-23-new-committer.md
@@ -0,0 +1,57 @@
+---
+title: 热烈欢迎 HertzBeat 小伙伴新晋社区 Committer!
+author: orangeCatDeveloper
+author_title: orangeCatDeveloper
+author_url: https://github.com/orangeCatDeveloper
+tags: [community]
+description: orangeCatDeveloper 分享从积压多年的老 issue
入手，修复数据可靠性、安全与部署问题并新增基础设施监控，成长为 Committer 的经历。
+keywords:
+ [
+ open source monitoring system,
+ alerting system,
+ AI observability,
+ Apache,
+ Apache Committer,
+ HertzBeat,
+ ]
+cover_headline: Welcome orangeCatDeveloper
+cover_kicker: New Committer
+---
+
+> 大家好，非常荣幸经 Apache HertzBeat PMC 投票，当选为 Apache HertzBeat™ Committer。
+
+## 个人介绍
+
+我是 [orangeCatDeveloper](https://github.com/orangeCatDeveloper)，一名专注于 AI
基础设施的软件工程师。我关心的是 AI 系统里不太被看见的那一半：给模型供数的数据管道，让 AI Agent
安全运行的执行环境，以及判断这一切是否运转正常的可观测性。
+
+## 初识 Apache HertzBeat
+
+第一次认真看 HertzBeat，是冲着它的 AI 能力去的。它内置了 AI 对话和 MCP
Server，大模型可以直接查监控数据、创建监控、分析告警。HertzBeat 是 Apache 软件基金会的顶级项目。截至 2026 年 9 月，它在
GitHub 上有 7400 多个 Star、300 多位贡献者，官方 Docker 镜像被拉取了 22 万多次。我在采集器里修掉的每一个
bug，都会随着新版本进入世界各地用户的监控系统。
+
+往下翻代码和 issue
之后，我的注意力反而落到了另一头。模型再聪明，读到的也只是采集器交上来的那些数字。数字错了，它会很认真地给出一个错误的结论。issue
列表里正好有不少这类问题，有的已经挂了两三年，我就从那里开始了。
+
+## 开源贡献之路
+
+截至 2026 年 9 月，我一共为 HertzBeat 合并了 [37 个
PR](https://github.com/search?q=is%3Apr+is%3Amerged+author%3AorangeCatDeveloper+repo%3Aapache%2Fhertzbeat+repo%3Aapache%2Fhertzbeat-helm-chart&type=pullrequests)，关掉了
28 个 issue，其中 19 个挂了一年以上，最早的一个提于 2023 年 1
月。采集器本质上就是一条数据管道：从一百多种系统里取数，交给告警、看板，现在还有大模型。我做的事，大多是让这条管道靠得住。
+
+最典型的是一个 2023 年就有人报告的问题：HertzBeat
一重启，网站监控就集体报慢，可网站本身好好的。根因是所有采集任务在重启后挤在同一秒触发，互相抢资源。我复现之后，给每个任务的首次调度加了随机抖动，让它们不再同时触发。在同样的测试负载下，监控上报的
p50 响应时间从 2865ms 降到 617ms，p95 从 6801ms 降到 1497ms。这类问题不报错，只是让数据悄悄失真，而告警、看板和 AI
分析读的都是这些数据。类似的还有编辑监控后历史数据被拆成多份、边界数据被悄悄丢弃等问题。
+
+安全是另一头。监控系统通常部署在内网深处，手里握着大量服务器和数据库的凭证，它自己不能成为最薄弱的一环。项目原先依赖的 Bouncy Castle 1.68
所在的版本线已经停更，带着 7 个已公开、且不会再被修复的 CVE，我把它迁移到了仍在维护、没有已知漏洞的
1.85。另外修了证书监控在关闭校验后仍然检查主机名的问题，并给邮件服务器加了证书校验开关，方便内网自签证书的场景。
+
+还有一些是部署上的问题。有一次数据库迁移和启动时的自动建表撞在了一起，全新安装会卡在启动阶段，之后再也起不来。我让迁移脚本可以重复执行，并让已经卡住的实例能自动恢复。在
Helm Chart 仓库，我修好了 CI，并为 Kubernetes 部署加上了启动探针。我还新增了 etcd 监控，它是 Kubernetes
控制面的核心存储。
+
+## 社区参与和成长
+
+这段时间最大的收获，是学会了怎么写一个让 reviewer 省力的
PR：先说用户遇到了什么，再说原因，最后贴修复前后的真实输出。大家的时间都很宝贵，证据摆在那里，讨论就会快很多。
+
+成为 Committer 之后，我想多花些时间在 review 上，帮新来的朋友把第一个 PR 顺利合进去。自己这边，接下来想做三件事：继续清理积压的老
issue；定期排查已经停更或带有已知漏洞的依赖，把供应链风险挡在发版之前；继续补齐基础设施层的监控覆盖，让 AI 分析拿到的数据覆盖得更全、也更可靠。
+
+## 给开源开发者的建议
+
+如果不知道从哪里开始，可以把 issue 列表按创建时间从旧到新翻一翻。老 issue
都是用户真实遇到过的问题，很多已经讨论到有方向了。先复现，有些会发现早就修好了，剩下的就值得花力气。修完把前后对比贴出来，这比任何解释都管用。
+
+AI 工具可以用，但提交的每一行代码，你都得能讲清楚为什么这样写。
+
+## 结语
+
+感谢社区各位对每个 PR 的耐心 review，也感谢 PMC 的提名和投票。HertzBeat 正在往 AI 驱动的可观测平台走，我希望它交给 AI
的每一个数字都经得起推敲。接下来继续一起加油。
diff --git
a/home/i18n/zh-cn/docusaurus-plugin-content-docs/current/help/alert_sms.md
b/home/i18n/zh-cn/docusaurus-plugin-content-docs/current/help/alert_sms.md
index b81d0c8afd..1e78ac000b 100644
--- a/home/i18n/zh-cn/docusaurus-plugin-content-docs/current/help/alert_sms.md
+++ b/home/i18n/zh-cn/docusaurus-plugin-content-docs/current/help/alert_sms.md
@@ -149,7 +149,7 @@ alerter:
### smslocal短信配置
-smslocal是一款面向企业的一体化短信服务平台，具备诸如多种发送方式、强大的安全性以及全天候支持等特性。你可以参考 smslocal
的[开发者文档](https://www.smslocal.com/developer/)来进行配置。
+smslocal是一款面向企业的一体化短信服务平台，具备诸如多种发送方式、强大的安全性以及全天候支持等特性。你可以参考 smslocal
的[开发者文档](https://www.smslocal.com/resources/docs/)来进行配置。
在 `application.yml` 中添加/填写以下 smslocal 配置内容（请用你自己的短信服务器配置参数替换相关参数）：
diff --git a/home/src/pages/team/data/member.json
b/home/src/pages/team/data/member.json
index 9577bf9f6e..0e03279c36 100644
--- a/home/src/pages/team/data/member.json
+++ b/home/src/pages/team/data/member.json
@@ -151,6 +151,11 @@
"githubId": "17263766",
"gitUrl": "https://github.com/delei";,
"name": "Delei"
+ },
+ {
+ "githubId": "95899648",
+ "gitUrl": "https://github.com/orangeCatDeveloper";,
+ "name": "orangeCatDeveloper"
}
]
}
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
