Support client dynamic registration

zdz · May 24, 2022 · bd23ccc · bd23ccc
1 parent 3763423
commit bd23ccc
Show file tree

Hide file tree

Showing 20 changed files with 899 additions and 310 deletions.
diff --git a/README.md b/README.md
@@ -8,14 +8,17 @@
 <img width="1215" alt="image" src="https://user-images.githubusercontent.com/152173/165957689-d35714a9-f7f8-49f7-9573-97d4cf3c2f79.png">
 <img width="1436" alt="image" src="https://user-images.githubusercontent.com/152173/165958225-25fc8fda-5798-42f8-bac5-72d778c0bab5.png">
 
+<h2>Table of Contents</h2>
+
 - [✨ Rust 版 ServerStatus 云探针](#-rust-版-serverstatus-云探针)
   - [1. 介绍](#1-介绍)
     - [🍀 主题](#-主题)
   - [2. 安装部署](#2-安装部署)
     - [2.1 快速体验](#21-快速体验)
-    - [2.2 服务管理脚本部署，感谢 @Colsro 提供](#22-服务管理脚本部署感谢-colsro-提供)
-    - [2.3 Railway 部署](#23-railway-部署)
-    - [2.4 前后端分离部署](#24-前后端分离部署)
+    - [2.2 快速部署](#22-快速部署)
+    - [2.3 服务管理脚本部署，感谢 @Colsro 提供](#23-服务管理脚本部署感谢-colsro-提供)
+    - [2.4 Railway 部署](#24-railway-部署)
+    - [2.5 前后端分离部署](#25-前后端分离部署)
   - [3. 服务端说明](#3-服务端说明)
     - [3.1 配置文件 `config.toml`](#31-配置文件-configtoml)
     - [3.2 服务端运行](#32-服务端运行)
@@ -38,15 +41,17 @@
 - 支持 `systemd` 开机自启
 - 其它功能，如 🗺️  见 [wiki](https://github.com/zdz/ServerStatus-Rust/wiki)
 
-演示：[tz-rust.vercel.app](https://tz-rust.vercel.app)
+演示：[ssr.rs](https://d.ssr.rs) | [vercel.app](https://tz-rust.vercel.app)
 |
 下载：[Releases](https://github.com/zdz/ServerStatus-Rust/releases)
 |
 反馈：[Discussions](https://github.com/zdz/ServerStatus-Rust/discussions)
 
+📕 完整文档迁移至 [doc.ssr.rs](https://doc.ssr.rs)
+
 ### 🍀 主题
 
-如果你觉得你修改的主题还不错，欢迎分享/PR，前端单独部署方法参见 [#37](https://github.com/zdz/ServerStatus-Rust/discussions/37)
+如果你觉得你创造/修改的主题还不错，欢迎分享/PR，前端单独部署方法参见 [#37](https://github.com/zdz/ServerStatus-Rust/discussions/37)
 
 <details>
   <summary>Hotaru 主题</summary>
@@ -70,7 +75,11 @@ bash -ex one-touch.sh
 # 自定义部署可参照 one-touch.sh 脚本
 ```
 
-### 2.2 服务管理脚本部署，感谢 [@Colsro](https://github.com/Colsro) 提供
+### 2.2 快速部署
+
+参见 [快速部署](https://doc.ssr.rs/rapid_deploy)
+
+### 2.3 服务管理脚本部署，感谢 [@Colsro](https://github.com/Colsro) 提供
 <details>
   <summary>管理脚本使用说明</summary>
 
@@ -115,14 +124,14 @@ help:
 </details>
 
 
-### 2.3 Railway 部署
+### 2.4 Railway 部署
 
 懒得配置 `Nginx`，`SSL` 证书？试试
 [在 Railway 部署 Server 教程](https://github.com/zdz/ServerStatus-Rust/wiki/Railway)
 
 [![Deploy on Railway](https://railway.app/button.svg)](https://railway.app/new/template/kzT46l?referralCode=pJYbdU)
 
-### 2.4 前后端分离部署
+### 2.5 前后端分离部署
 
 <details>
   <summary>前后端分离部署</summary>
@@ -195,15 +204,26 @@ admin_pass = ""
 # name 主机唯一标识，不可重复，alias 为展示名
 # 使用 ansible 批量部署时可以用主机 hostname 作为 name，统一密码
 # notify = false 单独禁止单台机器的告警，一般针对网络差，频繁上下线
-# monthstart = 1 没启用 vnstat 时，表示月流量从每月哪天开始统计
+# monthstart = 1 没启用vnstat时，表示月流量从每月哪天开始统计
 # disabled = true 单机禁用，跟删除这条配置的效果一样
 hosts = [
-  {name = "h1", password = "p1", alias = "n1", location = "🏠", type = "kvm"},
-  {name = "h2", password = "p2", alias = "n2", location = "🏢", type = "kvm", notify = true},
-  {name = "h3", password = "p3", alias = "n3", location = "🏝️", type = "kvm", monthstart = 1},
-  {name = "h4", password = "p4", alias = "n4", location = "🏢", type = "kvm", disabled = false},
+  {name = "h1", password = "p1", alias = "n1", location = "🏠", type = "kvm", notify = true},
+  {name = "h2", password = "p2", alias = "n2", location = "🏢", type = "kvm", disabled = false},
+  {name = "h3", password = "p3", alias = "n3", location = "🏡", type = "kvm", monthstart = 1},
 ]
 
+# 动态注册模式，不再需要针对每一个主机做单独配置
+# gid 为模板组id, 动态注册唯一标识，不可重复
+hosts_group = [
+  # 可以按国家地区或用途来做分组
+  {gid = "g1", password = "pp", location = "🏠", type = "kvm", notify = true},
+  {gid = "g2", password = "pp", location = "🏢", type = "kvm", notify = true},
+  # 例如不发送通知可以单独做一组
+  {gid = "silent", password = "pp", location = "🏡", type = "kvm", notify = false},
+]
+# 动态注册模式下，无效数据清理间隔，默认 30s
+group_gc = 30
+
 # 不开启告警，可忽略后面配置，或者删除不需要的通知方式
 # 告警间隔默认为30s
 notify_interval = 30
@@ -277,27 +297,31 @@ docker-compose up -d
 # rust client 可用参数
 ./stat_client -h
 OPTIONS:
-    -6, --ipv6             ipv6 only, default:false
-    -a, --addr <ADDR>      [default: http://127.0.0.1:8080/report]
-        --cm <CM_ADDR>     China Mobile probe addr [default: cm.tz.cloudcpp.com:80]
-        --ct <CT_ADDR>     China Telecom probe addr [default: ct.tz.cloudcpp.com:80]
-        --cu <CU_ADDR>     China Unicom probe addr [default: cu.tz.cloudcpp.com:80]
-        --disable-extra    disable extra info report, default:false
-        --disable-ping     disable ping, default:false
-        --disable-tupd     disable t/u/p/d, default:false
-    -h, --help             Print help information
-        --ip-info          show ip info, default:false
-        --json             use json protocol, default:false
-    -n, --vnstat           enable vnstat, default:false
-    -p, --pass <PASS>      password [default: p1]
-    -u, --user <USER>      username [default: h1]
-    -V, --version          Print version information
+    -6, --ipv6               ipv6 only, default:false
+    -a, --addr <ADDR>        [default: http://127.0.0.1:8080/report]
+        --alias <ALIAS>      alias for host [default: unknown]
+        --cm <CM_ADDR>       China Mobile probe addr [default: cm.tz.cloudcpp.com:80]
+        --ct <CT_ADDR>       China Telecom probe addr [default: ct.tz.cloudcpp.com:80]
+        --cu <CU_ADDR>       China Unicom probe addr [default: cu.tz.cloudcpp.com:80]
+        --disable-extra      disable extra info report, default:false
+        --disable-ping       disable ping, default:false
+        --disable-tupd       disable t/u/p/d, default:false
+    -g, --gid <GID>          group id [default: ]
+    -h, --help               Print help information
+        --ip-info            show ip info, default:false
+        --json               use json protocol, default:false
+    -n, --vnstat             enable vnstat, default:false
+    -p, --pass <PASS>        password [default: p1]
+    -u, --user <USER>        username [default: h1]
+    -V, --version            Print version information
+    -w, --weight <WEIGHT>    weight for rank [default: 0]
 
 # 一些参数说明
 --ip-info       # 显示本机ip信息后立即退出，目前使用 ip-api.com 数据
 --disable-extra # 不上报系统信息和IP信息
 --disable-ping  # 停用三网延时和丢包率探测
 --disable-tupd  # 不上报 tcp/udp/进程数/线程数，减少CPU占用
+-w, --weight    # 排序加分，微调让主机靠前显示，无强迫症可忽略
 ```
 
 ### 4.2 跨平台版本 (`Window`, `Linux`, `...`)
@@ -361,10 +385,6 @@ vnstat --version
 vnstat -m
 vnstat --json m
 
-# server config.toml 开启 vnstat
-# 从 v1.3.6 不再需要在 server 配置开启，client 自由选择启用与否，client 可部分打开，部分关闭
-vnstat = true
-
 # client 使用 -n 参数开启 vnstat 统计
 ./stat_client -a "grpc://127.0.0.1:9394" -u h1 -p p1 -n
 # 或
@@ -449,8 +469,8 @@ OPTIONS:
 <details>
   <summary>关于这个轮子</summary>
 
-  之前一直在使用 `Prometheus` + `Grafana` + `Alertmanager` + `node_exporter` 做VPS监控，这也是业界比较成熟的监控方案，用过一段时间后，发现非生产环境的话，很多监控指标都用不上，反而显得有些重。
-  而 `ServerStatus` 很好，足够简单和轻量，一眼可以看尽大好山河，只是 `c++` 版本很久没迭代过，自己的一些需求在原版上不是很好修改，如自带 `tcp` 上报对跨区机器不是很友好，也不方便对上报的链路优化 等等。过年的时候正值疫情闲来无事，学习 `Rust` 正好需要个小项目练手，于是撸了个 `ServerStatus` 来练手，项目后面会继续维护但不会增加复杂的功能，保持小而美，简单部署，配合 [Uptime Kuma](https://github.com/louislam/uptime-kuma) 基本上可以满足个人大部分监控需求。
+  之前一直在使用 `Prometheus` + `Grafana` + `Alertmanager` + `node_exporter` 做VPS监控，这也是业界比较成熟的监控方案，用过一段时间后，发现非生产环境，很多监控指标都用不上，反而显得有些重。
+  而 `ServerStatus` 很好，足够简单和轻量，一眼可以看尽所有小机机，只是 `c++` 版本很久没迭代过，自己的一些需求在原版上不是很好修改，如自带 `tcp` 上报对跨区机器不是很友好，也不方便对上报的链路做优化 等等。过年的时候正值疫情闲来无事，学习 `Rust` 正好需要个小项目练手，于是撸了个 `ServerStatus` 来练手，项目后面会继续维护但不会增加复杂的功能，保持小而美，简单部署，配合 [Uptime Kuma](https://github.com/louislam/uptime-kuma) 基本上可以满足个人大部分监控需求。
 
 </details>
 

diff --git a/client/Cargo.toml b/client/Cargo.toml
@@ -1,7 +1,7 @@
 [package]
 edition = "2021"
 name = "stat_client"
-version = "1.4.2"
+version = "1.5.0"
 
 rust-version = "1.60"
 
@@ -35,6 +35,7 @@ sysinfo = "0.23"
 tokio = {version = "1", features = ["full"]}
 tonic = {version = "0.7", features = ["tokio-rustls"]}
 tower = { version = "0.4" }
+md5 = "0.7.0"
 
 [features]
 default = ["native"]

diff --git a/client/src/grpc.rs b/client/src/grpc.rs
@@ -33,7 +33,16 @@ pub async fn report(args: &Args, stat_base: &mut StatRequest) -> anyhow::Result<
         );
     }
 
-    let token = MetadataValue::try_from(format!("{}@_@{}", args.user, args.pass))?;
+    let auth_user: String;
+    let ssr_auth: &[u8];
+    if args.gid.is_empty() {
+        auth_user = args.user.to_string();
+        ssr_auth = b"single";
+    } else {
+        auth_user = args.gid.to_string();
+        ssr_auth = b"group";
+    }
+    let token = MetadataValue::try_from(format!("{}@_@{}", auth_user, args.pass))?;
 
     let channel = Channel::from_shared(args.addr.to_string())?
         .connect()
@@ -43,6 +52,9 @@ pub async fn report(args: &Args, stat_base: &mut StatRequest) -> anyhow::Result<
     let grpc_client =
         ServerStatusClient::with_interceptor(timeout_channel, move |mut req: Request<()>| {
             req.metadata_mut().insert("authorization", token.clone());
+            req.metadata_mut()
+                .insert("ssr-auth", MetadataValue::try_from(ssr_auth).unwrap());
+
             Ok(req)
         });
 

diff --git a/client/src/main.rs b/client/src/main.rs
@@ -68,6 +68,15 @@ pub struct Args {
     json: bool,
     #[clap(short = '6', long = "ipv6", help = "ipv6 only, default:false")]
     ipv6: bool,
+    // #[clap(long = "debug", help = "debug mode, default:false")]
+    // debug: bool,
+    // for group
+    #[clap(short, long, default_value = "", help = "group id")]
+    gid: String,
+    #[clap(long = "alias", default_value = "unknown", help = "alias for host")]
+    alias: String,
+    #[clap(short, long, default_value = "0", help = "weight for rank")]
+    weight: u64,
 }
 
 fn sample_all(args: &Args, stat_base: &StatRequest) -> StatRequest {
@@ -145,8 +154,16 @@ fn http_report(args: &Args, stat_base: &mut StatRequest) -> Result<()> {
 
         let client = http_client.clone();
         let url = args.addr.to_string();
-        let auth_user = args.user.to_string();
         let auth_pass = args.pass.to_string();
+        let auth_user: String;
+        let ssr_auth: &str;
+        if args.gid.is_empty() {
+            auth_user = args.user.to_string();
+            ssr_auth = "single";
+        } else {
+            auth_user = args.gid.to_string();
+            ssr_auth = "group";
+        }
 
         // http
         tokio::spawn(async move {
@@ -155,6 +172,7 @@ fn http_report(args: &Args, stat_base: &mut StatRequest) -> Result<()> {
                 .basic_auth(auth_user, Some(auth_pass))
                 .timeout(Duration::from_secs(3))
                 .header(header::CONTENT_TYPE, content_type)
+                .header("ssr-auth", ssr_auth)
                 .body(body_data.unwrap())
                 .send()
                 .await
@@ -196,7 +214,7 @@ async fn refresh_ip_info(args: &Args) {
 #[tokio::main]
 async fn main() -> Result<()> {
     pretty_env_logger::init();
-    let args = Args::parse();
+    let mut args = Args::parse();
     dbg!(&args);
 
     if args.ip_info {
@@ -205,19 +223,21 @@ async fn main() -> Result<()> {
         process::exit(0);
     }
 
+    // support check
+    if !System::IS_SUPPORTED {
+        panic!("当前系统不支持，请切换到Python跨平台版本!");
+    }
+
     let sys_info = sys_info::collect_sys_info(&args);
     let sys_info_json = serde_json::to_string(&sys_info)?;
+    let sys_id = sys_info::gen_sys_id(&sys_info);
+    eprintln!("sys id: {}", sys_id);
     eprintln!("sys info: {}", sys_info_json);
 
     if let Ok(mut o) = G_CONFIG.lock() {
         o.sys_info = Some(sys_info);
     }
 
-    // support check
-    if !System::IS_SUPPORTED {
-        panic!("当前系统不支持，请切换到Python跨平台版本!");
-    }
-
     // use native
     #[cfg(all(feature = "native", not(feature = "sysinfo")))]
     {
@@ -250,8 +270,22 @@ async fn main() -> Result<()> {
         online4: ipv4,
         online6: ipv6,
         vnstat: args.vnstat,
+        weight: args.weight,
+        version: env!("CARGO_PKG_VERSION").to_string(),
         ..Default::default()
     };
+    if !args.gid.is_empty() {
+        stat_base.gid = args.gid.to_owned();
+        if stat_base.name.eq("h1") {
+            stat_base.name = sys_id;
+        }
+        if args.alias.eq("unknown") {
+            args.alias = stat_base.name.to_owned();
+        } else {
+            stat_base.alias = args.alias.to_owned();
+        }
+    }
+    // dbg!(&stat_base);
 
     if args.addr.starts_with("http") {
         let result = http_report(&args, &mut stat_base);

diff --git a/client/src/sys_info.rs b/client/src/sys_info.rs
@@ -83,6 +83,7 @@ pub fn start_net_speed_collect_t() {
     });
 }
 
+// TODO
 pub fn sample(args: &Args, stat: &mut StatRequest) {
     stat.version = env!("CARGO_PKG_VERSION").to_string();
     stat.vnstat = args.vnstat;
@@ -211,3 +212,19 @@ pub fn collect_sys_info(args: &Args) -> SysInfo {
 
     info_pb
 }
+
+pub fn gen_sys_id(sys_info: &SysInfo) -> String {
+    format!(
+        "{:x}",
+        md5::compute(format!(
+            "{}/{}/{}/{}/{}/{}/{}",
+            sys_info.host_name,
+            sys_info.os_name,
+            sys_info.os_arch,
+            sys_info.os_family,
+            sys_info.os_release,
+            sys_info.kernel_version,
+            sys_info.cpu_brand,
+        ))
+    )
+}
diff --git a/common/Cargo.toml b/common/Cargo.toml
@@ -1,7 +1,7 @@
 [package]
 edition = "2021"
 name = "stat_common"
-version = "1.0.0"
+version = "1.1.0"
 
 authors = ["doge <[email protected]>"]
 categories = ["monitoring-tools"]

diff --git a/common/proto/server_status.proto b/common/proto/server_status.proto
@@ -89,13 +89,16 @@ message StatRequest {
 
   optional SysInfo sys_info = 37;
   optional IpInfo ip_info = 38;
+
+  // group
+  string gid = 39;
+  string alias = 40;
+  uint64 weight = 41;
 }
 
 message Response {
   int32 code = 1;
   string message = 2;
 }
 
-service ServerStatus {
-  rpc Report(StatRequest) returns (Response);
-}
+service ServerStatus { rpc Report(StatRequest) returns (Response); }