⚡️ pprof 性能调优 / 计划任务支持秒级

.gitignore

@@ -10,8 +10,11 @@
 
 # Output of the go coverage tool, specifically when used with LiteIDE
 *.out
+*.pprof
 .idea
 /data
 /dist
 .DS_Store
 /main
+/cmd/agent/main
+/cmd/dashboard/main

README.md

@@ -4,7 +4,7 @@
   <br>
   <small><i>LOGO designed by <a href="https://xio.ng" target="_blank">熊大</a> .</i></small>
   <br><br>
-<img src="https://img.shields.io/github/workflow/status/naiba/nezha/Dashboard%20image?label=Dash%20v0.10.1&logo=github&style=for-the-badge">&nbsp;<img src="https://img.shields.io/github/v/release/naiba/nezha?color=brightgreen&label=Agent&style=for-the-badge&logo=github">&nbsp;<img src="https://img.shields.io/github/workflow/status/naiba/nezha/Agent%20release?label=Agent%20CI&logo=github&style=for-the-badge">&nbsp;<img src="https://img.shields.io/badge/Installer-v0.7.0-brightgreen?style=for-the-badge&logo=linux">
+<img src="https://img.shields.io/github/workflow/status/naiba/nezha/Dashboard%20image?label=Dash%20v0.10.2&logo=github&style=for-the-badge">&nbsp;<img src="https://img.shields.io/github/v/release/naiba/nezha?color=brightgreen&label=Agent&style=for-the-badge&logo=github">&nbsp;<img src="https://img.shields.io/github/workflow/status/naiba/nezha/Agent%20release?label=Agent%20CI&logo=github&style=for-the-badge">&nbsp;<img src="https://img.shields.io/badge/Installer-v0.7.0-brightgreen?style=for-the-badge&logo=linux">
   <br>
   <br>
   <p>:trollface: <b>哪吒监控</b> 一站式轻监控轻运维系统。支持系统状态、HTTP(SSL 证书变更、即将到期、到期)、TCP、Ping 监控报警，命令批量执行和计划任务。</p>
@@ -14,9 +14,9 @@
 
 \>> [我们的用户](https://www.google.com/search?q="powered+by+哪吒监控"&filter=0) (Google)
 
-| 默认主题                                                | DayNight [@JackieSung](https://github.com/JackieSung4ev) | hotaru                                                                 |
-| ------------------------------------------------------- | -------------------------------------------------------- | ---------------------------------------------------------------------- |
-| ![首页截图1](https://s3.ax1x.com/2020/12/07/DvTCwD.jpg) | <img src="https://s3.ax1x.com/2021/01/20/sfJv2q.jpg"/>   | <img src="https://s3.ax1x.com/2020/12/09/rPF4xJ.png" width="1600px" /> |
+| 默认主题                                                    | DayNight [@JackieSung](https://github.com/JackieSung4ev)     | hotaru                                                                     |
+| ----------------------------------------------------------- | ------------------------------------------------------------ | -------------------------------------------------------------------------- |
+| ![默认主题](resource/template/theme-default/screenshot.png) | ![daynight](resource/template/theme-daynight/screenshot.png) | <img src="resource/template/theme-hotaru/screenshot.png" width="1600px" /> |
 
 ## 安装脚本
 
@@ -36,11 +36,12 @@ CN=true sudo ./nezha.sh
 
 _\* 使用 WatchTower 可以自动更新面板，Windows 终端可以使用 nssm 配置自启动（见尾部教程）_
 
-### 特殊技能
+### 增强配置
 
 通过执行 `./nezha-agent --help` 查看支持的参数，如果你使用一键脚本，可以编辑 `/etc/systemd/system/nezha-agent.service`，在 `ExecStart=` 这一行的末尾加上
 
 - `--skip-conn` 不监控连接数，机场/连接密集型机器推荐设置，不然比较占 CPU([shirou/gopsutil/issues#220](https://github.com/shirou/gopsutil/issues/220))
+- `--skip-procs` 不监控进程数，也可以降低 agent 占用
 - `--disable-auto-update` 禁止 Agent 自动更新
 
 ## 功能说明
@@ -230,7 +231,7 @@ URL 里面也可放置占位符，请求时会进行简单的字符串替换。
 </details>
 
 <details>
-    <summary>Agent 不断重启/无法启动 ？</summary>
+    <summary>Agent不上线不启动自检流程</summary>
 
 1. 直接执行 `/opt/nezha/agent/nezha-agent -s 面板IP或非CDN域名:面板RPC端口 -p Agent密钥 -d` 查看日志是否是 DNS 问题。
 2. `nc -v 域名/IP 面板RPC端口` 或者 `telnet 域名/IP 面板RPC端口` 检验是否是网络问题，检查本机与面板服务器出入站防火墙，如果单机无法判断可借助 <https://port.ping.pe/> 提供的端口检查工具进行检测。
@@ -243,7 +244,7 @@ URL 里面也可放置占位符，请求时会进行简单的字符串替换。
 
 首先在 release 下载对应的二进制解压 tar.gz 包后放置到 `/root`，然后 `chmod +x /root/nezha-agent` 赋予执行权限，然后创建 `/etc/init.d/nezha-service`：
 
-```
+```shell
 #!/bin/sh /etc/rc.common
 
 START=99
@@ -272,8 +273,9 @@ restart() {
 </details>
 
 <details>
-    <summary>实时通道断开/终端连接失败</summary>
-使用反向代理时需要针对 `/ws` 路径的 WebSocket 进行特别配置以支持实时更新服务器状态。
+    <summary>实时通道断开/网页终端连接失败</summary>
+
+使用反向代理时需要针对 `/ws`,`/terminal` 路径的 WebSocket 进行特别配置以支持实时更新服务器状态和 **WebSSH**。
 
 - Nginx(宝塔)：在你的 nginx 配置文件中加入以下代码
 
@@ -285,7 +287,6 @@ restart() {
 
       location ~ ^/(ws|terminal/.+)$  {
           proxy_pass http://ip:站点访问端口;
-          proxy_http_version 1.1;
           proxy_set_header Upgrade $http_upgrade;
           proxy_set_header Connection "Upgrade";
           proxy_set_header Host $host;
@@ -295,6 +296,15 @@ restart() {
   }
   ```
 
+  如果非宝塔，还要在 `server{}` 中添加上这一段
+
+  ```nginx
+  location / {
+    proxy_pass http://ip:站点访问端口;
+    proxy_set_header Host $host;
+  }
+  ```
+
 - CaddyServer v1（v2 无需特别配置）
 
   ```Caddyfile

cmd/agent/main.go

@@ -30,25 +30,21 @@ import (
 
 type AgentConfig struct {
 	SkipConnectionCount bool
+	SkipProcsCount      bool
 	DisableAutoUpdate   bool
 	Debug               bool
 	Server              string
 	ClientSecret        string
 }
 
-func init() {
-	http.DefaultClient.Timeout = time.Second * 5
-	flag.CommandLine.ParseErrorsWhitelist.UnknownFlags = true
-}
-
 var (
-	version   string
-	agentConf AgentConfig
+	version string
+	client  pb.NezhaServiceClient
+	inited  bool
 )
 
 var (
-	client     pb.NezhaServiceClient
-	inited     bool
+	agentConf  AgentConfig
 	updateCh   = make(chan struct{}) // Agent 自动更新间隔
 	httpClient = &http.Client{
 		CheckRedirect: func(req *http.Request, via []*http.Request) error {
@@ -63,6 +59,11 @@ const (
 	networkTimeOut = time.Second * 5  // 普通网络超时
 )
 
+func init() {
+	http.DefaultClient.Timeout = time.Second * 5
+	flag.CommandLine.ParseErrorsWhitelist.UnknownFlags = true
+}
+
 func main() {
 	// 来自于 GoReleaser 的版本号
 	monitor.Version = version
@@ -71,6 +72,7 @@ func main() {
 	flag.StringVarP(&agentConf.Server, "server", "s", "localhost:5555", "管理面板RPC端口")
 	flag.StringVarP(&agentConf.ClientSecret, "password", "p", "", "Agent连接Secret")
 	flag.BoolVar(&agentConf.SkipConnectionCount, "skip-conn", false, "不监控连接数")
+	flag.BoolVar(&agentConf.SkipProcsCount, "skip-procs", false, "不监控进程数")
 	flag.BoolVar(&agentConf.DisableAutoUpdate, "disable-auto-update", false, "禁用自动升级")
 	flag.Parse()
 
@@ -88,7 +90,6 @@ func run() {
 	}
 
 	go pty.DownloadDependency()
-
 	// 上报服务器信息
 	go reportState()
 	// 更新IP信息
@@ -168,8 +169,8 @@ func receiveTasks(tasks pb.NezhaService_RequestTaskClient) error {
 		}
 		go func() {
 			defer func() {
-				if recover() != nil {
-					println("task panic", task)
+				if err := recover(); err != nil {
+					println("task panic", task, err)
 				}
 			}()
 			doTask(task)
@@ -209,7 +210,7 @@ func reportState() {
 		if client != nil && inited {
 			monitor.TrackNetworkSpeed()
 			timeOutCtx, cancel := context.WithTimeout(context.Background(), networkTimeOut)
-			_, err = client.ReportSystemState(timeOutCtx, monitor.GetState(agentConf.SkipConnectionCount).PB())
+			_, err = client.ReportSystemState(timeOutCtx, monitor.GetState(agentConf.SkipConnectionCount, agentConf.SkipProcsCount).PB())
 			cancel()
 			if err != nil {
 				println("reportState error", err)
@@ -229,7 +230,7 @@ func doSelfUpdate() {
 	println("检查更新：", v)
 	latest, err := selfupdate.UpdateSelf(v, "naiba/nezha")
 	if err != nil {
-		println("自动更新失败：", err)
+		println("更新失败：", err)
 		return
 	}
 	if !latest.Version.Equals(v) {
@@ -282,7 +283,7 @@ func handleHttpGetTask(task *pb.Task, result *pb.TaskResult) {
 	}
 	if err == nil {
 		// 检查 SSL 证书信息
-		if len(resp.TLS.PeerCertificates) > 0 {
+		if resp.TLS != nil && len(resp.TLS.PeerCertificates) > 0 {
 			c := resp.TLS.PeerCertificates[0]
 			result.Data = c.Issuer.CommonName + "|" + c.NotAfter.In(time.Local).String()
 		}

cmd/agent/monitor/monitor.go

@@ -5,7 +5,6 @@ import (
 	"regexp"
 	"runtime"
 	"strings"
-	"sync/atomic"
 	"syscall"
 	"time"
 
@@ -15,20 +14,27 @@ import (
 	"github.com/shirou/gopsutil/v3/load"
 	"github.com/shirou/gopsutil/v3/mem"
 	"github.com/shirou/gopsutil/v3/net"
+	"github.com/shirou/gopsutil/v3/process"
 
 	"github.com/naiba/nezha/model"
 )
 
-var Version string = "debug"
-var netInSpeed, netOutSpeed, netInTransfer, netOutTransfer, lastUpdate uint64
-var expectDiskFsTypes = []string{
-	"apfs", "ext4", "ext3", "ext2", "f2fs", "reiserfs", "jfs", "btrfs",
-	"fuseblk", "zfs", "simfs", "ntfs", "fat32", "exfat", "xfs", "fuse.rclone",
-}
-var excludeNetInterfaces = []string{
-	"lo", "tun", "docker", "veth", "br-", "vmbr", "vnet", "kube",
-}
-var getMacDiskNo = regexp.MustCompile(`\/dev\/disk(\d)s.*`)
+var (
+	Version           string = "debug"
+	expectDiskFsTypes        = []string{
+		"apfs", "ext4", "ext3", "ext2", "f2fs", "reiserfs", "jfs", "btrfs",
+		"fuseblk", "zfs", "simfs", "ntfs", "fat32", "exfat", "xfs", "fuse.rclone",
+	}
+	excludeNetInterfaces = []string{
+		"lo", "tun", "docker", "veth", "br-", "vmbr", "vnet", "kube",
+	}
+	getMacDiskNo = regexp.MustCompile(`\/dev\/disk(\d)s.*`)
+)
+
+var (
+	netInSpeed, netOutSpeed, netInTransfer, netOutTransfer, lastUpdateNetStats uint64
+	cachedBootTime                                                             time.Time
+)
 
 func GetHost() *model.Host {
 	hi, _ := host.Info()
@@ -58,6 +64,8 @@ func GetHost() *model.Host {
 		swapMemTotal = mv.SwapTotal
 	}
 
+	cachedBootTime = time.Now().Add(time.Duration(-1 * int64(hi.BootTime*1000)))
+
 	return &model.Host{
 		Platform:        hi.OS,
 		PlatformVersion: hi.PlatformVersion,
@@ -74,8 +82,12 @@ func GetHost() *model.Host {
 	}
 }
 
-func GetState(skipConnectionCount bool) *model.HostState {
-	hi, _ := host.Info()
+func GetState(skipConnectionCount bool, skipProcsCount bool) *model.HostState {
+	var procs []int32
+	if !skipProcsCount {
+		procs, _ = process.Pids()
+	}
+
 	mv, _ := mem.VirtualMemory()
 
 	var swapMemUsed uint64
@@ -92,11 +104,11 @@ func GetState(skipConnectionCount bool) *model.HostState {
 	if err == nil {
 		cpuPercent = cp[0]
 	}
+
 	_, diskUsed := getDiskTotalAndUsed()
 	loadStat, _ := load.Avg()
 
 	var tcpConnCount, udpConnCount uint64
-
 	if !skipConnectionCount {
 		conns, _ := net.Connections("all")
 		for i := 0; i < len(conns); i++ {
@@ -114,17 +126,17 @@ func GetState(skipConnectionCount bool) *model.HostState {
 		MemUsed:        mv.Total - mv.Available,
 		SwapUsed:       swapMemUsed,
 		DiskUsed:       diskUsed,
-		NetInTransfer:  atomic.LoadUint64(&netInTransfer),
-		NetOutTransfer: atomic.LoadUint64(&netOutTransfer),
-		NetInSpeed:     atomic.LoadUint64(&netInSpeed),
-		NetOutSpeed:    atomic.LoadUint64(&netOutSpeed),
-		Uptime:         hi.Uptime,
+		NetInTransfer:  netInTransfer,
+		NetOutTransfer: netOutTransfer,
+		NetInSpeed:     netInSpeed,
+		NetOutSpeed:    netOutSpeed,
+		Uptime:         uint64(time.Since(cachedBootTime).Seconds()),
 		Load1:          loadStat.Load1,
 		Load5:          loadStat.Load5,
 		Load15:         loadStat.Load15,
 		TcpConnCount:   tcpConnCount,
 		UdpConnCount:   udpConnCount,
-		ProcessCount:   hi.Procs,
+		ProcessCount:   uint64(len(procs)),
 	}
 }
 
@@ -140,14 +152,14 @@ func TrackNetworkSpeed() {
 			innerNetOutTransfer += v.BytesSent
 		}
 		now := uint64(time.Now().Unix())
-		diff := now - atomic.LoadUint64(&lastUpdate)
+		diff := now - lastUpdateNetStats
 		if diff > 0 {
-			atomic.StoreUint64(&netInSpeed, (innerNetInTransfer-atomic.LoadUint64(&netInTransfer))/diff)
-			atomic.StoreUint64(&netOutSpeed, (innerNetOutTransfer-atomic.LoadUint64(&netOutTransfer))/diff)
+			netInSpeed = (innerNetInTransfer - netInTransfer) / diff
+			netOutSpeed = (innerNetOutTransfer - netOutTransfer) / diff
 		}
-		atomic.StoreUint64(&netInTransfer, innerNetInTransfer)
-		atomic.StoreUint64(&netOutTransfer, innerNetOutTransfer)
-		atomic.StoreUint64(&lastUpdate, now)
+		netInTransfer = innerNetInTransfer
+		netOutTransfer = innerNetOutTransfer
+		lastUpdateNetStats = now
 	}
 }
 

cmd/agent/monitor/myip.go

@@ -35,13 +35,21 @@ func UpdateIP() {
 	for {
 		ipv4 := fetchGeoIP(geoIPApiList, false)
 		ipv6 := fetchGeoIP(geoIPApiList, true)
-		cachedIP = fmt.Sprintf("ip(v4:%s,v6:[%s])", ipv4.IP, ipv6.IP)
+		if ipv4.IP == "" && ipv6.IP == "" {
+			time.Sleep(time.Minute)
+			continue
+		}
+		if ipv4.IP == "" || ipv6.IP == "" {
+			cachedIP = fmt.Sprintf("%s%s", ipv4.IP, ipv6.IP)
+		} else {
+			cachedIP = fmt.Sprintf("%s/%s", ipv4.IP, ipv6.IP)
+		}
 		if ipv4.CountryCode != "" {
 			cachedCountry = ipv4.CountryCode
 		} else if ipv6.CountryCode != "" {
 			cachedCountry = ipv6.CountryCode
 		}
-		time.Sleep(time.Minute * 10)
+		time.Sleep(time.Minute * 30)
 	}
 }
 

cmd/dashboard/controller/controller.go

@@ -112,7 +112,7 @@ func ServeWeb(port uint) *http.Server {
 		},
 	})
 	r.Static("/static", "resource/static")
-	r.LoadHTMLGlob("resource/template/**/*")
+	r.LoadHTMLGlob("resource/template/**/*.html")
 	routers(r)
 
 	page404 := func(c *gin.Context) {

cmd/dashboard/controller/member_api.go

@@ -267,19 +267,22 @@ func (ma *memberAPI) addOrEditCron(c *gin.Context) {
 		cr.Cover = cf.Cover
 		err = json.Unmarshal([]byte(cf.ServersRaw), &cr.Servers)
 	}
-	if err == nil {
-		_, err = cron.ParseStandard(cr.Scheduler)
-	}
+	tx := dao.DB.Begin()
 	if err == nil {
 		if cf.ID == 0 {
-			err = dao.DB.Create(&cr).Error
+			err = tx.Create(&cr).Error
 		} else {
-			err = dao.DB.Save(&cr).Error
+			err = tx.Save(&cr).Error
 		}
 	}
 	if err == nil {
 		cr.CronID, err = dao.Cron.AddFunc(cr.Scheduler, dao.CronTrigger(cr))
 	}
+	if err == nil {
+		err = tx.Commit().Error
+	} else {
+		tx.Rollback()
+	}
 	if err != nil {
 		c.JSON(http.StatusOK, model.Response{
 			Code:    http.StatusBadRequest,