开发一款IIS自动检测修复工具

         

目录

实现的功能

技术栈与依赖 DLL

实现细节

变量

初始化操作

自定义cpu阈值

检测IIS应用程序池

获取自定义阈值

获取某个应用程序池的占用率

获取性能计数器实例名

Kill 并重新启动应用池

写入日志到 Log 目录，并显示在文本框中

实际运行效果

---

        此工具可放在服务器中，每隔3秒检测一次iis中所有应用程序池中异常的应用程序。

实现的功能

• 每3秒检测所有应用程序池中占用cpu过高的程序，cpu阈值可自行设置，超过阈值自动kill进程并且进行重启
• 每3秒检测iis应用程序池中有崩溃、异常停止的应用程序，且进行自动3次的重启，3次以后不会自动重启，且过滤掉手动停止的iis应用程序池。

此界面左边会显示所有在iis中的应用程序池的信息，右边则是textBox，持续滚动显示日志，图片出现异常是我的本地电脑没有部署IIS。

技术栈与依赖 DLL

类型	技术 / DLL	用途	路径说明
平台	WinForms (.NET Framework 4.5.2)	主程序框架	—
核心 DLL	Microsoft.Web.Administration.dll	操作 IIS 应用池状态与回收	C:\Windows\System32\inetsrv\
系统库	System.Diagnostics	获取进程信息与 CPU 使用率	.NET 自带
日志	System.IO	记录操作日志到本地文件	.NET 自带

⚠️ 注意：程序必须以 管理员权限运行 才能访问 IIS 的相关管理权限。

实现细节

• 变量

 private System.Threading.Timer MonitorTimer; //定义计时器private const int MonitorInterval = 3000; // 检测间隔：5秒一检测private string LogDirectory = Path.Combine(AppDomain.CurrentDomain.BaseDirectory, "Log");//日志目录private string CpuConfigPath = Path.Combine(AppDomain.CurrentDomain.BaseDirectory, "cpu_threshold.txt");//cpu阈值文件private Dictionary<string, int> RestartDict = new Dictionary<string, int>(); //应用程序池重启次数                                                                                              // 定义文件路径（可放在程序目录）private string RestartCountPath = Path.Combine(AppDomain.CurrentDomain.BaseDirectory, "restart_counts.json");//应用程序池重启次数文件private volatile bool mIsChecking = false;//标志位避免重入，防止资源竞争。(检测 IIS 应用池状态 CheckAppPools)private bool mIsClose = false;

• 初始化操作

/// <summary>
/// 初始化列表控件
/// </summary>
private void InitListView()
{listView1.View = View.Details;listView1.Columns.Add("应用程序池", 100);listView1.Columns.Add("进程ID", 100);listView1.Columns.Add("状态", 120);listView1.Columns.Add("CPU 使用率(%)", 120);// 启用 OwnerDrawlistView1.OwnerDraw = true;listView1.DrawColumnHeader += listView1_DrawColumnHeader;listView1.DrawSubItem += listView1_DrawSubItem;
}private void listView1_DrawColumnHeader(object sender, DrawListViewColumnHeaderEventArgs e)
{e.DrawDefault = true; // 默认绘制列头
}private void listView1_DrawSubItem(object sender, DrawListViewSubItemEventArgs e)
{if (e.ColumnIndex == 3) // 第四列是 cpuUsage{float cpu;int mCpuValue = GetCpuThreshold();if (float.TryParse(e.SubItem.Text, out cpu) && cpu >=mCpuValue) // 大于阈值时标红{e.Graphics.DrawString(e.SubItem.Text, listView1.Font, Brushes.Red, e.Bounds);}else{e.Graphics.DrawString(e.SubItem.Text, listView1.Font, Brushes.Black, e.Bounds);}}else{e.DrawDefault = true;}
}/// <summary>
/// 启动定时器监控
/// </summary>
private void StartMonitoring()
{MonitorTimer = new System.Threading.Timer(CheckAppPools, null, 0, MonitorInterval);
}

• 自定义cpu阈值

 if (button1.Text == "设置 CPU 阈值"){txtCpuThreshold.Visible = true;txtCpuThreshold.Text = GetCpuThreshold().ToString();button1.Text = "保存";label1.Location = new System.Drawing.Point(258, 12);}else{label1.Location = new System.Drawing.Point(129, 11);string input = txtCpuThreshold.Text.Trim();if (int.TryParse(input, out int threshold)){string path = Path.Combine(AppDomain.CurrentDomain.BaseDirectory, "cpu_threshold.txt");File.WriteAllText(path, threshold.ToString());txtCpuThreshold.Visible = false;button1.Text = "设置 CPU 阈值";label1.Text = "当前CPU阈值为：" + input + "%";MessageBox.Show($"CPU阈值已设置为 {threshold}%", "提示", MessageBoxButtons.OK, MessageBoxIcon.Information);}else{MessageBox.Show("请输入有效的整数值作为阈值。", "错误", MessageBoxButtons.OK, MessageBoxIcon.Error);}}

• 检测IIS应用程序池

private void CheckAppPools(object state)
{if (mIsChecking) return;mIsChecking = true;try{using (ServerManager serverManager = new ServerManager()){var mCurrentNames = new HashSet<string>();foreach (var pool in serverManager.ApplicationPools){string mProcessID = "";//进程IDstring name = pool.Name;string statusEn = pool.State.ToString(); //应用程序池状态string statusCn = GetStatusDescription(pool.State);//应用程序池状态中文描述double cpuUsage = GetCpuUsageByAppPool(name);//获取某个应用程序池的cpu使用率var wp = pool.WorkerProcesses.FirstOrDefault();if (wp != null){var process = Process.GetProcessById(wp.ProcessId);mProcessID=process.Id.ToString();}else{mProcessID = "NULL";}int threshold = GetCpuThreshold();//我们设置的阈值if (cpuUsage >= threshold){KillAndRestart(pool);//超过了阈值以后直接杀掉然后重启Log($"[{name}](进程ID:{mProcessID}) CPU 使用率过高 {cpuUsage}%，已杀掉该进程并且重启。状态：{statusCn}",true);}//如果是已停止的状态我们通过日志去查是不是崩溃导致的，避免是手动停止的一直重启造成崩溃if (statusEn == "Stopped"){string logSource = "WAS"; // 只筛选管理应用程序池（AppPool）的生命周期，包括启动、停止、崩溃等日志。string logName = "System"; //系统日志DateTime checkTime = DateTime.Now.AddMinutes(-5); // 查最近5分钟bool isCrash = false;try{EventLog eventLog = new EventLog(logName);foreach (EventLogEntry entry in eventLog.Entries){if (entry.TimeGenerated >= checkTime &&entry.Source == logSource &&entry.EntryType == EventLogEntryType.Error &&entry.Message.Contains(name)) // 判断是否提到目标AppPool{isCrash = true;Log($"[{name}](进程ID:{mProcessID})发现相关崩溃日志，准备尝试自动重启。奔溃日志：" +entry.Message,true);break;}}}catch (Exception ex){Log($"[{name}] 检查WAS事件日志失败：{ex.Message}",false);}// 重启次数限制（防止死循环）string restartKey = $"RestartCount_{name}";int restartCount = RestartDict.ContainsKey(restartKey) ? RestartDict[restartKey] : 0;int restartLimit = 3;if (isCrash){if (restartCount < restartLimit){KillAndRestart(pool);restartCount++;RestartDict[restartKey] = restartCount;File.WriteAllText(RestartCountPath, JsonConvert.SerializeObject(RestartDict, Formatting.Indented));Log($"[{name}] (进程ID:{mProcessID})状态异常，尝试 Kill 并重启。（当前次数：{restartCount}）",true);}else{Log($"[{name}](进程ID:{mProcessID}) 已达重启上限（{restartLimit}次），不再重启，防止死循环。",true);}}else{Log($"[{name}] (进程ID:{mProcessID})状态为 Stopped，但无崩溃迹象，疑似人工操作，跳过自动重启。", true);}}mCurrentNames.Add(name);if (!name.Contains("DefaultAppPool") && !name.Contains(".NET v4.5 Classic") && !name.Contains(".NET v4.5")){Invoke(new Action(() =>{var existingItem = listView1.Items.Cast<ListViewItem>().FirstOrDefault(i => i.Text == name);if (existingItem != null){existingItem.SubItems[1].Text = mProcessID;existingItem.SubItems[2].Text = statusEn + "（" + statusCn + "）";existingItem.SubItems[3].Text = cpuUsage.ToString("F2");}else{var item = new ListViewItem(new[]{name,mProcessID,statusEn + "（" + statusCn + "）",cpuUsage.ToString("F2")});listView1.Items.Add(item);}}));}}// 移除已不存在的项Invoke(new Action(() =>{for (int i = listView1.Items.Count - 1; i >= 0; i--){string name = listView1.Items[i].Text;if (!mCurrentNames.Contains(name)){listView1.Items.RemoveAt(i);}}}));}}catch (Exception ex){Log("检测IIS应用池出现异常: " + ex.Message,true);}finally{mIsChecking = false;}
}

• 获取自定义阈值

  private int GetCpuThreshold(){try{if (File.Exists(CpuConfigPath)){string text = File.ReadAllText(CpuConfigPath).Trim();if (int.TryParse(text, out int value))return value;}}catch { }return 70; // 默认值}

• 获取某个应用程序池的占用率

 private double GetCpuUsageByAppPool(string appPoolName){// 排除三个应用程序池名称if (appPoolName.Contains("DefaultAppPool") ||appPoolName.Contains(".NET v4.5 Classic") ||appPoolName.Contains(".NET v4.5")){//Log($"[{appPoolName}] 属于排除监控列表，跳过 CPU 监测。");return 0;}using (ServerManager manager = new ServerManager()){var pool = manager.ApplicationPools[appPoolName];if (pool == null){Log($"[{appPoolName}]该应用池对象为 null，跳过处理。",false);return 0;}var wp = pool.WorkerProcesses.FirstOrDefault();if (wp == null){Log($"[{appPoolName}]的WorkerProcesses(进程总和)count为0，可能尚未启动，跳过处理。",false);return 0;}try{var process = Process.GetProcessById(wp.ProcessId);// 获取唯一的性能计数器实例名（如 w3wp#1）进程名都是w3wp，所以要取一下IDstring instanceName = GetProcessInstanceName(appPoolName, process.Id);//Log($"[{appPoolName}] 对应进程名称：{process.ProcessName}，PID：{process.Id}，性能计数器实例名：{instanceName}");var counter = new PerformanceCounter("Process", "% Processor Time", instanceName, true);counter.NextValue(); // 第一次调用为初始化，不可信Thread.Sleep(2000); // 等待间隔至少 1 秒以上float rawValue = counter.NextValue();int cpuCount = Environment.ProcessorCount;float usage = rawValue / cpuCount;//Windows 的 "% Processor Time" 是所有 CPU 核心总和，例如 4 核 CPU，一个进程满载可能是 400 %//所以：如果rawValue = 200（代表该进程用掉了 200 %）//cpuCount = 4 假设cpu核数是4核//usage = 200 / 4 =50//这样才能和设置的阈值（例如80 %）做准确对比。Log($"[{appPoolName}]CPU使用率：{usage:F2}%（原始值：{rawValue:F2}%，核数：{cpuCount}）",false);return usage;}catch (Exception ex){Log($"[{appPoolName}] 获取 CPU 使用率失败：{ex.Message}",false);return 0;}}}

• 获取性能计数器实例名

private string GetProcessInstanceName(string appPoolName,int pid)
{try{PerformanceCounterCategory category = new PerformanceCounterCategory("Process");string[] instances = category.GetInstanceNames();foreach (string instance in instances){using (PerformanceCounter counter = new PerformanceCounter("Process", "ID Process", instance, true)){if ((int)counter.RawValue == pid){return instance;}}}}catch (Exception ex){Log("["+appPoolName+"]未找到 PID 为 "+pid+" 的性能计数器实例名",false);}return "";
}

• Kill 并重新启动应用池

private void KillAndRestart(ApplicationPool pool)
{string poolName = pool.Name;try{foreach (var wp in pool.WorkerProcesses){try { Process.GetProcessById(wp.ProcessId).Kill(); } catch { Log("kill并重新启动[" + poolName + "] 池时，没有找到"+ wp.ProcessId,true); }}pool.Stop();Thread.Sleep(1000);pool.Start();Log("[" + poolName + "]重启成功!",true);}catch (Exception ex){Log("["+poolName+"]重启失败: " + ex.Message,true);}
}

• 写入日志到 Log 目录，并显示在文本框中

private void Log(string message,bool mIsWriteTxt)
{try{if (mIsWriteTxt){if (!Directory.Exists(LogDirectory))Directory.CreateDirectory(LogDirectory);string mFileName = DateTime.Now.ToString("yyyyMMdd") + "Log.txt";string mFilePath = Path.Combine(LogDirectory, mFileName);string mLine = $"[{DateTime.Now:yyyy-MM-dd HH:mm:ss}] {message}";using (StreamWriter writer = new StreamWriter(mFilePath, true)){writer.WriteLine(mLine + Environment.NewLine);}}Invoke(new Action(() =>{// 判断当前文本框行数是否超过 200 行int currentLineCount = txtLog.Lines.Length;if (currentLineCount >= 200){txtLog.Clear(); // 超过行数则清空}string mLine = $"[{DateTime.Now:yyyy-MM-dd HH:mm:ss}] {message}";txtLog.AppendText(mLine + Environment.NewLine);}));}catch { }
}

实际运行效果

• 工具部署在多台服务器的桌面系统上，稳定运行数月。

• 显著减少了手工处理 IIS 挂死的情况，节省了近 80% 的运维时间。

• 在 CPU 高频波动的高峰期，应用池自动回收机制保证了服务稳定不崩溃。

• 日志功能也为问题回溯与性能分析提供了有力支持。