Metatron — Intranet Site Navigator

English | 中文

Why — pain points this solves

Team pain point	What Metatron does
Internal sites scattered — consoles, Grafana, CI, docs bookmarks live in every browser	One searchable page: category tabs, environment tags (prod/dev/intranet/SaaS), keyword search
Credentials passed around in chat	Per-site account/password with one-click copy (keep it behind an intranet-only ingress)
New hires don't know where anything is	A self-describing inventory with descriptions and auth notes per site
Maintaining a portal shouldn't need a build pipeline	Edit one HTML file → merge → Pods roll automatically (kustomize ConfigMap hash)

Features

• One HTML file is the whole site (manifests/site-nav.html): inline CSS/JS, zero external assets.
• Kustomize ConfigMap rollouts: configMapGenerator turns the HTML / nginx.conf into content-hashed ConfigMaps — edit the file, merge, and Pods roll automatically (ArgoCD friendly).
• Stock nginx:alpine: nothing to build or push; pull from anywhere.
• Ingress-ready: internal ALB + TLS example included.

Layout

manifests/
  site-nav.html      the whole site (data list + UI in one file)
  nginx.conf         nginx site config (charset, gzip, SPA fallback, /healthz)
  deployment.yaml    2 replicas of nginx:alpine, ConfigMap volumes
  service.yaml       ClusterIP :80
  ingress.yaml       internal ALB ingress example
  kustomization.yaml configMapGenerator wiring
argocd/
  app-metatron.yaml  ArgoCD Application (auto sync + prune + self-heal)

Run it

kubectl create ns metatron
kubectl apply -k manifests/

Or via ArgoCD: edit repoURL/targetRevision in argocd/app-metatron.yaml, then kubectl apply -f argocd/app-metatron.yaml.

Customizing

All site entries live in the SITES array inside site-nav.html:

{ name: "Grafana", url: "https://grafana.example.com/", cat: "observe",
  env: "aliyun", auth: "aliyun-ram", tags: ["dashboards"], desc: "..." }

Fields: cat (category tab), env (aliyun / intranet / saas), auth, tags, desc, and optional account/password for entries that need standalone credentials. Adjust to fit your own inventory.

⚠️ If you keep credentials in the data list, host the portal somewhere access-controlled (intranet-only ingress). A public copy should scrub them.

---

中文

English | 中文

为什么做这个(解决的痛点)

团队痛点	Metatron 的做法
内网站点散落各处——云控制台、Grafana、CI、文档收藏散在每个人的浏览器里	一页收拢:分类页签、环境标签(prod/dev/内网/SaaS)、关键词搜索
账号密码靠聊天工具传来传去	站点级 account/password 字段,一键复制(务必只挂内网 ingress)
新人不知道从哪进门	每个站点带描述与登录方式说明的自描述清单
维护一个导航台不该引入构建流水线	改一个 HTML 文件 → 合并 → Pod 自动滚动(kustomize ConfigMap 哈希)

特性

• 一个 HTML 就是整站(manifests/site-nav.html):内联 CSS/JS,零外部资源。
• Kustomize ConfigMap 滚动:用 configMapGenerator 把 HTML / nginx.conf 变成带内容哈希的 ConfigMap——改文件、合并,Pod 自动滚动更新(对 ArgoCD 友好)。
• 原生 nginx:alpine:不用构建不用推镜像,哪里都能拉。
• Ingress 就绪:附内网 ALB + TLS 的示例配置。

目录结构

manifests/
  site-nav.html      整站(数据清单 + UI 都在一个文件里)
  nginx.conf         nginx 站点配置(charset、gzip、SPA 回退、/healthz)
  deployment.yaml    双副本 nginx:alpine,挂 ConfigMap 卷
  service.yaml       ClusterIP :80
  ingress.yaml       内网 ALB ingress 示例
  kustomization.yaml configMapGenerator 接线
argocd/
  app-metatron.yaml  ArgoCD Application(自动同步 + prune + self-heal)

部署

kubectl create ns metatron
kubectl apply -k manifests/

或走 ArgoCD:改 argocd/app-metatron.yaml 里的 repoURL/targetRevision,然后 kubectl apply -f argocd/app-metatron.yaml。

定制

所有站点条目都在 site-nav.html 的 SITES 数组里:

{ name: "Grafana", url: "https://grafana.example.com/", cat: "observe",
  env: "aliyun", auth: "aliyun-ram", tags: ["大盘"], desc: "..." }

字段:cat(分类页签)、env(aliyun / intranet / saas)、auth、tags、desc,需要独立账号的条目可加 account/password。按自己的清单调整即可。

⚠️ 如果把凭据放在数据清单里,门户务必部署在访问受控的位置(仅内网 ingress);公开副本应先清掉凭据。