pm2-perfmonitor 2.6.1 → 2.6.3

package/lib/app.js

@@ -1,581 +1,581 @@
-const pmx = require('pmx')
-const pm2 = require('pm2')
-
-const { listAppsAsync, restartAppAsync } = require('./pm2-extra')
-const {
-  parseParamToArray,
-  parseParamToNumber,
-  parseBool,
-  safeToFixed,
-  sleepAsync,
-  getSysCpuUsageByPid,
-} = require('./utils')
-const { defaultOptions } = require('./defaults')
-const { sendMessage } = require('./message')
-const { performPerfSampling } = require('./perf-sampler')
-const { sendAlert } = require('./alert')
-const { getJobConfHostName } = require('./job-conf')
-const { isZombieCpuProcess, isZombieStateProcess } = require('./zombie-check')
-
-/**
- * @type { defaultOptions }
- */
-const conf = pmx.initModule({}, (err, incomingConf) => {
-  if (err) {
-    console.error(`[${incomingConf.module_name}] init module error:`, err)
-    process.exit(2)
-  }
-
-  return {
-    ...defaultOptions,
-    ...incomingConf,
-  }
-})
-
-const Probe = pmx.probe()
-const MODULE_NAME = conf.module_name
-const MODULE_ENABLED = parseBool(conf.enabled)
-const WORKER_INTERVAL = parseParamToNumber(conf.workerInterval)
-const INCLUDE_APPS = parseParamToArray(conf.includeApps)
-const EXCLUDE_APPS = parseParamToArray(conf.excludeApps)
-
-// zombie conf
-const ZOMBIE_DETECTION = parseBool(conf.zombieDetection)
-const AUTO_RESTART_WHEN_ZOMBIE_DETECTED = parseBool(
-  conf.autoRestartWhenZombieDetected,
-)
-const ZOMBIE_MAX_HITS = parseParamToNumber(conf.zombieMaxHits)
-const ZOMBIE_MAX_RESTARTS = parseParamToNumber(conf.zombieMaxRestarts)
-const zombieProcessDetectionStrategy = conf.zombieProcessDetectionStrategy
-
-// cpu conf
-const cpuOverloadDetection = parseBool(conf.cpuOverloadDetection)
-const cpuOverloadThreshold = parseParamToNumber(conf.cpuOverloadThreshold)
-const cpuOverloadMaxHits = parseParamToNumber(conf.cpuOverloadMaxHits)
-const enablePerfCollection = parseBool(conf.enablePerfCollection)
-const perfReportGenerationDir = conf.perfReportGenerationDir
-const flamegraphDir = conf.flamegraphDir
-const perfSampleDuration = parseParamToNumber(conf.perfSampleDuration)
-const perfSampleFrequency = parseParamToNumber(conf.perfSampleFrequency)
-const enableNodeInspectorCollection = parseBool(
-  conf.enableNodeInspectorCollection,
-)
-const nodeInspectorSampleDuration = parseParamToNumber(
-  conf.nodeInspectorSampleDuration,
-)
-
-// alert conf
-const enableAlert = parseBool(conf.enableAlert)
-const alertCmdPath = conf.alertCmdPath
-const alertEnv = conf.alertEnv
-const alertLevel = conf.alertLevel
-const jobHostNameConfPath = conf.jobHostNameConfPath
-
-// 存储每个进程的 CPU 采样历史（pm_id -> [cpu1, cpu2, ...]）
-const zombieCpuHistory = new Map()
-const zombieRestartHistory = new Map()
-const zombieRestartFailedHistory = new Map()
-
-const cpuOverloadHistory = new Map()
-const cpuOverloadRestartHistory = new Map()
-const cpuOverloadRestartFailedHistory = new Map()
-
-/**
- * @type { Map<number, number> }
- */
-const cpuUsageFromPM2 = new Map()
-/**
- * @type { Map<number, number> }
- */
-const cpuUsageFromSys = new Map()
-
-let isProcessCheckerRunning = false
-
-/**
- * perf 样本是否采集中
- * @type { Map<number,boolean> }
- */
-const perfSamplingStats = new Map()
-
-/**
- * @param {'log' | 'info' | 'error' | 'warn'} type
- *
- */
-const logger = (type, ...args) => {
-  return console[type](`[${MODULE_NAME}]`, ...args)
-}
-
-/**
- * 判断是否为僵尸进程：最近 ZOMBIE_MAX_HITS 次全是 0%
- * @param { number } pid
- * @param { number[] } cpus
- */
-const isZombieProcess = async (pid, cpus) => {
-  if (zombieProcessDetectionStrategy === 'zombie-state') {
-    const res = await isZombieStateProcess(pid)
-
-    return res.isZombie
-  } else if (zombieProcessDetectionStrategy === 'zero-cpu-consecutive') {
-    return isZombieCpuProcess({
-      cpus,
-      maxHits: ZOMBIE_MAX_HITS,
-    })
-  } else if (zombieProcessDetectionStrategy === 'zombie-state-and-zero-cpu') {
-    const { isZombie: zombieState, failed } = await isZombieStateProcess(pid)
-
-    const zombieCpu = isZombieCpuProcess({
-      cpus,
-      maxHits: ZOMBIE_MAX_HITS,
-    })
-
-    // 获取系统进程状态失败时，回退为通过 CPU 持续 0% 判定
-    if (failed) {
-      return zombieCpu
-    }
-
-    return zombieState && zombieCpu
-  } else if (zombieProcessDetectionStrategy === 'zombie-state-or-zero-cpu') {
-    const { isZombie: zombieState } = await isZombieStateProcess(pid)
-
-    const zombieCpu = isZombieCpuProcess({
-      cpus,
-      maxHits: ZOMBIE_MAX_HITS,
-    })
-
-    return zombieState || zombieCpu
-  }
-
-  return isZombieCpuProcess({
-    cpus,
-    maxHits: ZOMBIE_MAX_HITS,
-  })
-}
-
-/**
- * @param { number[] } cpus
- */
-const isCpuOverload = (cpus) => {
-  return (
-    cpus.length >= cpuOverloadMaxHits &&
-    cpus.every((v) => v >= cpuOverloadThreshold)
-  )
-}
-
-/**
- * @param { number } pm_id
- * @param { number  } appCpuUsage
- * @returns { number[] } 对应 pm_id 的 CPU 使用率数组
- */
-const setZombieCpuHistory = (pm_id, appCpuUsage) => {
-  if (!zombieCpuHistory.has(pm_id)) {
-    zombieCpuHistory.set(pm_id, [])
-  }
-
-  const history = zombieCpuHistory.get(pm_id)
-
-  history.push(appCpuUsage)
-
-  // 只保留最近 ZOMBIE_MAX_HITS 次记录
-  if (history.length > ZOMBIE_MAX_HITS) {
-    history.shift()
-  }
-
-  return history
-}
-
-/**
- * @param { number } pm_id
- * @param { number  } appCpuUsage
- * @returns { number[] } 对应 pm_id 的 CPU 使用率数组
- */
-const setCpuOverloadHistory = (pm_id, appCpuUsage) => {
-  if (!cpuOverloadHistory.has(pm_id)) {
-    cpuOverloadHistory.set(pm_id, [])
-  }
-
-  const history = cpuOverloadHistory.get(pm_id)
-
-  history.push(appCpuUsage)
-
-  // 只保留最近 x 次记录
-  if (history.length > cpuOverloadMaxHits) {
-    history.shift()
-  }
-
-  return history
-}
-
-const setZombieRestartFailedHistory = (pm_id) => {
-  if (!zombieRestartFailedHistory.has(pm_id)) {
-    zombieRestartFailedHistory.set(pm_id, 1)
-  } else {
-    zombieRestartFailedHistory.set(
-      pm_id,
-      zombieRestartFailedHistory.get(pm_id) + 1,
-    )
-  }
-}
-
-const setCpuOverloadRestartFailedHistory = (pm_id) => {
-  if (!cpuOverloadRestartFailedHistory.has(pm_id)) {
-    cpuOverloadRestartFailedHistory.set(pm_id, 1)
-  } else {
-    cpuOverloadRestartFailedHistory.set(
-      pm_id,
-      cpuOverloadRestartFailedHistory.get(pm_id) + 1,
-    )
-  }
-}
-
-/**
- * 发送重启警告
- * @param {string} title
- * @param {string} message
- */
-const sendRestartAlert = async (title, message) => {
-  if (!enableAlert) return
-
-  const datetime = new Date().toLocaleString()
-  const jobHostName = getJobConfHostName(jobHostNameConfPath)
-
-  return await sendAlert({
-    cmd: alertCmdPath,
-    env: alertEnv,
-    level: alertLevel,
-    title: `[${MODULE_NAME}] Alert: ${title}`,
-    content: `[${jobHostName}] [${datetime}] - ${message}`,
-  })
-}
-
-/**
- * check process
- */
-const processChecker = async () => {
-  if (isProcessCheckerRunning) return
-
-  try {
-    isProcessCheckerRunning = true
-
-    const apps = await listAppsAsync()
-
-    cpuUsageFromPM2.clear()
-    cpuUsageFromSys.clear()
-
-    const cpus = apps.map((v) => v.monit?.cpu)
-
-    const hasAbnormalMonitoringData = cpus.every((v) => typeof v !== 'number')
-
-    if (hasAbnormalMonitoringData) {
-      const details = apps.map((v) => {
-        const { name, pid, pm_id, monit } = v
-
-        return {
-          name,
-          pid,
-          pm_id,
-          monit,
-        }
-      })
-
-      logger('warn', 'Abnormal monitoring data exists: ', details)
-    }
-
-    for (const app of apps) {
-      const { name, pid, pm_id, monit, pm2_env } = app
-
-      const appStatus = pm2_env?.status
-
-      // 非目标应用，跳过
-      if (
-        MODULE_NAME === name ||
-        (INCLUDE_APPS.length > 0 && !INCLUDE_APPS.includes(name)) ||
-        (EXCLUDE_APPS.length > 0 && EXCLUDE_APPS.includes(name))
-      ) {
-        continue
-      }
-
-      // 只处理 online 状态的进程
-      if (appStatus !== 'online') {
-        // 进程不在 online 状态时，清空其历史记录，避免干扰
-        zombieCpuHistory.delete(pm_id)
-        cpuOverloadHistory.delete(pm_id)
-
-        continue
-      }
-
-      const pm2CpuUsage = monit?.cpu
-      const sysCpuUsage = await getSysCpuUsageByPid(pid)
-      const appCpuUsage =
-        typeof sysCpuUsage === 'number' ? sysCpuUsage : pm2CpuUsage
-
-      cpuUsageFromPM2.set(pm_id, pm2CpuUsage)
-      cpuUsageFromSys.set(pm_id, sysCpuUsage)
-
-      const cpuHistory = setZombieCpuHistory(pm_id, appCpuUsage)
-      const cpuHistory2 = setCpuOverloadHistory(pm_id, appCpuUsage)
-
-      // 发送消息通知对应应用进程，采样 CPU 性能
-      if (enableNodeInspectorCollection) {
-        if (appCpuUsage >= cpuOverloadThreshold) {
-          await sendMessage(pm_id, 'cpu-profile-start')
-          await sleepAsync(nodeInspectorSampleDuration * 1000)
-          await sendMessage(pm_id, 'cpu-profile-stop')
-        }
-      }
-
-      const zombieDetected = await isZombieProcess(pid, cpuHistory)
-
-      // 判断是否为僵尸：最近 ZOMBIE_MAX_HITS 次全是 0%
-      if (ZOMBIE_DETECTION && zombieDetected) {
-        logger(
-          'info',
-          `Zombie detected: ${name} (pm_id: ${pm_id}, pid: ${app.pid})`,
-        )
-
-        if (AUTO_RESTART_WHEN_ZOMBIE_DETECTED) {
-          if (
-            ZOMBIE_MAX_RESTARTS > 0 &&
-            zombieRestartHistory.get(pm_id) >= ZOMBIE_MAX_RESTARTS
-          ) {
-            continue
-          }
-
-          logger('info', 'restarting...')
-
-          try {
-            await restartAppAsync(pm_id)
-
-            if (!zombieRestartHistory.has(pm_id)) {
-              zombieRestartHistory.set(pm_id, 1)
-            } else {
-              const history = zombieRestartHistory.get(pm_id)
-
-              zombieRestartHistory.set(pm_id, history + 1)
-            }
-
-            logger(
-              'info',
-              `[ZOMBIE] Restarted ${name} (pm_id: ${pm_id}) successfully!!! Restarted ${zombieRestartHistory.get(pm_id)} times`,
-            )
-
-            await sendRestartAlert(
-              `The zombie process has been restarted!`,
-              `appName: ${name}, pid: ${pid}, pm_id: ${pm_id}, restarted: ${zombieRestartHistory.get(pm_id)} times`,
-            )
-
-            // 重启后清除该进程的历史记录，避免刚重启又被判定为僵尸
-            zombieCpuHistory.delete(pm_id)
-          } catch (restartErr) {
-            logger(
-              'error',
-              `[ZOMBIE] Restart failed for ${name} (pm_id: ${pm_id}):`,
-              restartErr,
-            )
-
-            setZombieRestartFailedHistory(pm_id)
-          }
-        }
-      }
-      // CPU 是否持续过载
-      else if (cpuOverloadDetection && isCpuOverload(cpuHistory2)) {
-        logger(
-          'info',
-          `CPU Overload detected: ${name} (pm_id: ${pm_id}, pid: ${app.pid})`,
-        )
-
-        if (enablePerfCollection && !perfSamplingStats.get(pm_id)) {
-          perfSamplingStats.set(pm_id, true)
-
-          await performPerfSampling({
-            pid,
-            moduleName: MODULE_NAME,
-            perfDir: perfReportGenerationDir,
-            flamegraphDir,
-            sampleDuration: perfSampleDuration,
-            sampleFrequency: perfSampleFrequency,
-          })
-
-          perfSamplingStats.delete(pm_id)
-        }
-
-        try {
-          logger('info', 'restarting...')
-
-          await restartAppAsync(pm_id)
-
-          if (!cpuOverloadRestartHistory.has(pm_id)) {
-            cpuOverloadRestartHistory.set(pm_id, 1)
-          } else {
-            cpuOverloadRestartHistory.set(
-              pm_id,
-              cpuOverloadRestartHistory.get(pm_id) + 1,
-            )
-          }
-
-          logger(
-            'info',
-            `[CPU OVERLOAD] Restarted ${name} (pm_id: ${pm_id}) successfully!!! Restarted ${cpuOverloadRestartHistory.get(pm_id)} times`,
-          )
-
-          await sendRestartAlert(
-            `CPU overload process restarted!`,
-            `appName: ${name}, pid: ${pid}, pm_id: ${pm_id}, restarted: ${cpuOverloadRestartHistory.get(pm_id)} times`,
-          )
-
-          cpuOverloadHistory.delete(pm_id)
-        } catch (restartErr) {
-          logger(
-            'error',
-            `[CPU OVERLOAD] Restart failed for ${name} (pm_id: ${pm_id}):`,
-            restartErr,
-          )
-
-          setCpuOverloadRestartFailedHistory(pm_id)
-        }
-      }
-    }
-  } catch (err) {
-    logger('error', err)
-  } finally {
-    isProcessCheckerRunning = false
-  }
-}
-
-const runModule = () => {
-  if (!MODULE_ENABLED) return
-
-  // connect to local pm2
-  pm2.connect((err) => {
-    if (err) {
-      logger('error', `PM2 connection error:`, err)
-
-      process.exit(1)
-    }
-
-    logger('info', 'Connected to PM2, starting monitor...')
-
-    processChecker()
-
-    setInterval(() => {
-      processChecker()
-    }, WORKER_INTERVAL)
-  })
-
-  /** PROB PMX **/
-  Probe.metric({
-    name: 'Zombie Restarts',
-    value: () => {
-      const res = []
-
-      for (const [k, v] of zombieRestartHistory) {
-        if (v > 0) {
-          res.push([k, v])
-        }
-      }
-
-      if (!res.length) return 'N/A'
-
-      return res.map((v) => `[${v[0]}]:${v[1]}`).join(';')
-    },
-  })
-
-  Probe.metric({
-    name: 'Zombie Restarts (failed)',
-    value: () => {
-      const res = []
-
-      for (const [k, v] of zombieRestartFailedHistory) {
-        if (v > 0) {
-          res.push([k, v])
-        }
-      }
-
-      if (!res.length) return 'N/A'
-
-      return res.map((v) => `[${v[0]}]:${v[1]}`).join(';')
-    },
-  })
-
-  Probe.metric({
-    name: 'CPU Overload Restarts',
-    value: () => {
-      const res = []
-
-      for (const [k, v] of cpuOverloadRestartHistory) {
-        if (v > 0) {
-          res.push([k, v])
-        }
-      }
-
-      if (!res.length) return 'N/A'
-
-      return res.map((v) => `[${v[0]}]:${v[1]}`).join(';')
-    },
-  })
-
-  Probe.metric({
-    name: 'CPU Overload Restarts (failed)',
-    value: () => {
-      const res = []
-
-      for (const [k, v] of cpuOverloadRestartFailedHistory) {
-        if (v > 0) {
-          res.push([k, v])
-        }
-      }
-
-      if (!res.length) return 'N/A'
-
-      return res.map((v) => `[${v[0]}]:${v[1]}`).join(';')
-    },
-  })
-
-  Probe.metric({
-    name: 'Processes in Sampling (perf)',
-    value: () => {
-      const res = []
-
-      for (const [k, v] of perfSamplingStats) {
-        if (v === true) {
-          res.push(k)
-        }
-      }
-
-      if (!res.length) return 'N/A'
-
-      return res.join(', ')
-    },
-  })
-
-  Probe.metric({
-    name: 'CPU (pm2)',
-    value: () => {
-      const res = []
-
-      for (const [k, v] of cpuUsageFromPM2) {
-        res.push([k, v])
-      }
-
-      if (!res.length) return 'N/A'
-
-      return res.map((v) => `[${v[0]}]:${v[1]}%`).join(';')
-    },
-  })
-
-  Probe.metric({
-    name: 'CPU (sys)',
-    value: () => {
-      const res = []
-
-      for (const [k, v] of cpuUsageFromSys) {
-        res.push([k, v])
-      }
-
-      if (!res.length) return 'N/A'
-
-      return res.map((v) => `[${v[0]}]:${v[1]}%`).join(';')
-    },
-  })
-}
-
-runModule()
+const pmx = require('pmx')
+const pm2 = require('pm2')
+
+const { listAppsAsync, restartAppAsync } = require('./pm2-extra')
+const {
+  parseParamToArray,
+  parseParamToNumber,
+  parseBool,
+  safeToFixed,
+  sleepAsync,
+  getSysCpuUsageByPid,
+} = require('./utils')
+const { defaultOptions } = require('./defaults')
+const { sendMessage } = require('./message')
+const { performPerfSampling } = require('./perf-sampler')
+const { sendAlert } = require('./alert')
+const { getJobConfHostName } = require('./job-conf')
+const { isZombieCpuProcess, isZombieStateProcess } = require('./zombie-check')
+
+/**
+ * @type { defaultOptions }
+ */
+const conf = pmx.initModule({}, (err, incomingConf) => {
+  if (err) {
+    console.error(`[${incomingConf.module_name}] init module error:`, err)
+    process.exit(2)
+  }
+
+  return {
+    ...defaultOptions,
+    ...incomingConf,
+  }
+})
+
+const Probe = pmx.probe()
+const MODULE_NAME = conf.module_name
+const MODULE_ENABLED = parseBool(conf.enabled)
+const WORKER_INTERVAL = parseParamToNumber(conf.workerInterval)
+const INCLUDE_APPS = parseParamToArray(conf.includeApps)
+const EXCLUDE_APPS = parseParamToArray(conf.excludeApps)
+
+// zombie conf
+const ZOMBIE_DETECTION = parseBool(conf.zombieDetection)
+const AUTO_RESTART_WHEN_ZOMBIE_DETECTED = parseBool(
+  conf.autoRestartWhenZombieDetected,
+)
+const ZOMBIE_MAX_HITS = parseParamToNumber(conf.zombieMaxHits)
+const ZOMBIE_MAX_RESTARTS = parseParamToNumber(conf.zombieMaxRestarts)
+const zombieProcessDetectionStrategy = conf.zombieProcessDetectionStrategy
+
+// cpu conf
+const cpuOverloadDetection = parseBool(conf.cpuOverloadDetection)
+const cpuOverloadThreshold = parseParamToNumber(conf.cpuOverloadThreshold)
+const cpuOverloadMaxHits = parseParamToNumber(conf.cpuOverloadMaxHits)
+const enablePerfCollection = parseBool(conf.enablePerfCollection)
+const perfReportGenerationDir = conf.perfReportGenerationDir
+const flamegraphDir = conf.flamegraphDir
+const perfSampleDuration = parseParamToNumber(conf.perfSampleDuration)
+const perfSampleFrequency = parseParamToNumber(conf.perfSampleFrequency)
+const enableNodeInspectorCollection = parseBool(
+  conf.enableNodeInspectorCollection,
+)
+const nodeInspectorSampleDuration = parseParamToNumber(
+  conf.nodeInspectorSampleDuration,
+)
+
+// alert conf
+const enableAlert = parseBool(conf.enableAlert)
+const alertCmdPath = conf.alertCmdPath
+const alertEnv = conf.alertEnv
+const alertLevel = conf.alertLevel
+const jobHostNameConfPath = conf.jobHostNameConfPath
+
+// 存储每个进程的 CPU 采样历史（pm_id -> [cpu1, cpu2, ...]）
+const zombieCpuHistory = new Map()
+const zombieRestartHistory = new Map()
+const zombieRestartFailedHistory = new Map()
+
+const cpuOverloadHistory = new Map()
+const cpuOverloadRestartHistory = new Map()
+const cpuOverloadRestartFailedHistory = new Map()
+
+/**
+ * @type { Map<number, number> }
+ */
+const cpuUsageFromPM2 = new Map()
+/**
+ * @type { Map<number, number> }
+ */
+const cpuUsageFromSys = new Map()
+
+let isProcessCheckerRunning = false
+
+/**
+ * perf 样本是否采集中
+ * @type { Map<number,boolean> }
+ */
+const perfSamplingStats = new Map()
+
+/**
+ * @param {'log' | 'info' | 'error' | 'warn'} type
+ *
+ */
+const logger = (type, ...args) => {
+  return console[type](`[${MODULE_NAME}]`, ...args)
+}
+
+/**
+ * 判断是否为僵尸进程：最近 ZOMBIE_MAX_HITS 次全是 0%
+ * @param { number } pid
+ * @param { number[] } cpus
+ */
+const isZombieProcess = async (pid, cpus) => {
+  if (zombieProcessDetectionStrategy === 'zombie-state') {
+    const res = await isZombieStateProcess(pid)
+
+    return res.isZombie
+  } else if (zombieProcessDetectionStrategy === 'zero-cpu-consecutive') {
+    return isZombieCpuProcess({
+      cpus,
+      maxHits: ZOMBIE_MAX_HITS,
+    })
+  } else if (zombieProcessDetectionStrategy === 'zombie-state-and-zero-cpu') {
+    const { isZombie: zombieState, failed } = await isZombieStateProcess(pid)
+
+    const zombieCpu = isZombieCpuProcess({
+      cpus,
+      maxHits: ZOMBIE_MAX_HITS,
+    })
+
+    // 获取系统进程状态失败时，回退为通过 CPU 持续 0% 判定
+    if (failed) {
+      return zombieCpu
+    }
+
+    return zombieState && zombieCpu
+  } else if (zombieProcessDetectionStrategy === 'zombie-state-or-zero-cpu') {
+    const { isZombie: zombieState } = await isZombieStateProcess(pid)
+
+    const zombieCpu = isZombieCpuProcess({
+      cpus,
+      maxHits: ZOMBIE_MAX_HITS,
+    })
+
+    return zombieState || zombieCpu
+  }
+
+  return isZombieCpuProcess({
+    cpus,
+    maxHits: ZOMBIE_MAX_HITS,
+  })
+}
+
+/**
+ * @param { number[] } cpus
+ */
+const isCpuOverload = (cpus) => {
+  return (
+    cpus.length >= cpuOverloadMaxHits &&
+    cpus.every((v) => v >= cpuOverloadThreshold)
+  )
+}
+
+/**
+ * @param { number } pm_id
+ * @param { number  } appCpuUsage
+ * @returns { number[] } 对应 pm_id 的 CPU 使用率数组
+ */
+const setZombieCpuHistory = (pm_id, appCpuUsage) => {
+  if (!zombieCpuHistory.has(pm_id)) {
+    zombieCpuHistory.set(pm_id, [])
+  }
+
+  const history = zombieCpuHistory.get(pm_id)
+
+  history.push(appCpuUsage)
+
+  // 只保留最近 ZOMBIE_MAX_HITS 次记录
+  if (history.length > ZOMBIE_MAX_HITS) {
+    history.shift()
+  }
+
+  return history
+}
+
+/**
+ * @param { number } pm_id
+ * @param { number  } appCpuUsage
+ * @returns { number[] } 对应 pm_id 的 CPU 使用率数组
+ */
+const setCpuOverloadHistory = (pm_id, appCpuUsage) => {
+  if (!cpuOverloadHistory.has(pm_id)) {
+    cpuOverloadHistory.set(pm_id, [])
+  }
+
+  const history = cpuOverloadHistory.get(pm_id)
+
+  history.push(appCpuUsage)
+
+  // 只保留最近 x 次记录
+  if (history.length > cpuOverloadMaxHits) {
+    history.shift()
+  }
+
+  return history
+}
+
+const setZombieRestartFailedHistory = (pm_id) => {
+  if (!zombieRestartFailedHistory.has(pm_id)) {
+    zombieRestartFailedHistory.set(pm_id, 1)
+  } else {
+    zombieRestartFailedHistory.set(
+      pm_id,
+      zombieRestartFailedHistory.get(pm_id) + 1,
+    )
+  }
+}
+
+const setCpuOverloadRestartFailedHistory = (pm_id) => {
+  if (!cpuOverloadRestartFailedHistory.has(pm_id)) {
+    cpuOverloadRestartFailedHistory.set(pm_id, 1)
+  } else {
+    cpuOverloadRestartFailedHistory.set(
+      pm_id,
+      cpuOverloadRestartFailedHistory.get(pm_id) + 1,
+    )
+  }
+}
+
+/**
+ * 发送重启警告
+ * @param {string} title
+ * @param {string} message
+ */
+const sendRestartAlert = async (title, message) => {
+  if (!enableAlert) return
+
+  const datetime = new Date().toLocaleString()
+  const jobHostName = getJobConfHostName(jobHostNameConfPath)
+
+  return await sendAlert({
+    cmd: alertCmdPath,
+    env: alertEnv,
+    level: alertLevel,
+    title: `[${MODULE_NAME}] Alert: ${title}`,
+    content: `[${jobHostName}] [${datetime}] - ${message}`,
+  })
+}
+
+/**
+ * check process
+ */
+const processChecker = async () => {
+  if (isProcessCheckerRunning) return
+
+  try {
+    isProcessCheckerRunning = true
+
+    const apps = await listAppsAsync()
+
+    cpuUsageFromPM2.clear()
+    cpuUsageFromSys.clear()
+
+    const cpus = apps.map((v) => v.monit?.cpu)
+
+    const hasAbnormalMonitoringData = cpus.every((v) => typeof v !== 'number')
+
+    if (hasAbnormalMonitoringData) {
+      const details = apps.map((v) => {
+        const { name, pid, pm_id, monit } = v
+
+        return {
+          name,
+          pid,
+          pm_id,
+          monit,
+        }
+      })
+
+      logger('warn', 'Abnormal monitoring data exists: ', details)
+    }
+
+    for (const app of apps) {
+      const { name, pid, pm_id, monit, pm2_env } = app
+
+      const appStatus = pm2_env?.status
+
+      // 非目标应用，跳过
+      if (
+        MODULE_NAME === name ||
+        (INCLUDE_APPS.length > 0 && !INCLUDE_APPS.includes(name)) ||
+        (EXCLUDE_APPS.length > 0 && EXCLUDE_APPS.includes(name))
+      ) {
+        continue
+      }
+
+      // 只处理 online 状态的进程
+      if (appStatus !== 'online') {
+        // 进程不在 online 状态时，清空其历史记录，避免干扰
+        zombieCpuHistory.delete(pm_id)
+        cpuOverloadHistory.delete(pm_id)
+
+        continue
+      }
+
+      const pm2CpuUsage = monit?.cpu
+      const sysCpuUsage = await getSysCpuUsageByPid(pid)
+      const appCpuUsage =
+        typeof sysCpuUsage === 'number' ? sysCpuUsage : pm2CpuUsage
+
+      cpuUsageFromPM2.set(pm_id, pm2CpuUsage)
+      cpuUsageFromSys.set(pm_id, sysCpuUsage)
+
+      const cpuHistory = setZombieCpuHistory(pm_id, appCpuUsage)
+      const cpuHistory2 = setCpuOverloadHistory(pm_id, appCpuUsage)
+
+      // 发送消息通知对应应用进程，采样 CPU 性能
+      if (enableNodeInspectorCollection) {
+        if (appCpuUsage >= cpuOverloadThreshold) {
+          await sendMessage(pm_id, 'cpu-profile-start')
+          await sleepAsync(nodeInspectorSampleDuration * 1000)
+          await sendMessage(pm_id, 'cpu-profile-stop')
+        }
+      }
+
+      const zombieDetected = await isZombieProcess(pid, cpuHistory)
+
+      // 判断是否为僵尸：最近 ZOMBIE_MAX_HITS 次全是 0%
+      if (ZOMBIE_DETECTION && zombieDetected) {
+        logger(
+          'info',
+          `Zombie detected: ${name} (pm_id: ${pm_id}, pid: ${app.pid})`,
+        )
+
+        if (AUTO_RESTART_WHEN_ZOMBIE_DETECTED) {
+          if (
+            ZOMBIE_MAX_RESTARTS > 0 &&
+            zombieRestartHistory.get(pm_id) >= ZOMBIE_MAX_RESTARTS
+          ) {
+            continue
+          }
+
+          logger('info', 'restarting...')
+
+          try {
+            await restartAppAsync(pm_id)
+
+            if (!zombieRestartHistory.has(pm_id)) {
+              zombieRestartHistory.set(pm_id, 1)
+            } else {
+              const history = zombieRestartHistory.get(pm_id)
+
+              zombieRestartHistory.set(pm_id, history + 1)
+            }
+
+            logger(
+              'info',
+              `[ZOMBIE] Restarted ${name} (pm_id: ${pm_id}) successfully!!! Restarted ${zombieRestartHistory.get(pm_id)} times`,
+            )
+
+            await sendRestartAlert(
+              `The zombie process has been restarted!`,
+              `appName: ${name}, pid: ${pid}, pm_id: ${pm_id}, restarted: ${zombieRestartHistory.get(pm_id)} times`,
+            )
+
+            // 重启后清除该进程的历史记录，避免刚重启又被判定为僵尸
+            zombieCpuHistory.delete(pm_id)
+          } catch (restartErr) {
+            logger(
+              'error',
+              `[ZOMBIE] Restart failed for ${name} (pm_id: ${pm_id}):`,
+              restartErr,
+            )
+
+            setZombieRestartFailedHistory(pm_id)
+          }
+        }
+      }
+      // CPU 是否持续过载
+      else if (cpuOverloadDetection && isCpuOverload(cpuHistory2)) {
+        logger(
+          'info',
+          `CPU Overload detected: ${name} (pm_id: ${pm_id}, pid: ${app.pid})`,
+        )
+
+        if (enablePerfCollection && !perfSamplingStats.get(pm_id)) {
+          perfSamplingStats.set(pm_id, true)
+
+          await performPerfSampling({
+            pid,
+            moduleName: MODULE_NAME,
+            perfDir: perfReportGenerationDir,
+            flamegraphDir,
+            sampleDuration: perfSampleDuration,
+            sampleFrequency: perfSampleFrequency,
+          })
+
+          perfSamplingStats.delete(pm_id)
+        }
+
+        try {
+          logger('info', 'restarting...')
+
+          await restartAppAsync(pm_id)
+
+          if (!cpuOverloadRestartHistory.has(pm_id)) {
+            cpuOverloadRestartHistory.set(pm_id, 1)
+          } else {
+            cpuOverloadRestartHistory.set(
+              pm_id,
+              cpuOverloadRestartHistory.get(pm_id) + 1,
+            )
+          }
+
+          logger(
+            'info',
+            `[CPU OVERLOAD] Restarted ${name} (pm_id: ${pm_id}) successfully!!! Restarted ${cpuOverloadRestartHistory.get(pm_id)} times`,
+          )
+
+          await sendRestartAlert(
+            `CPU overload process restarted!`,
+            `appName: ${name}, pid: ${pid}, pm_id: ${pm_id}, restarted: ${cpuOverloadRestartHistory.get(pm_id)} times`,
+          )
+
+          cpuOverloadHistory.delete(pm_id)
+        } catch (restartErr) {
+          logger(
+            'error',
+            `[CPU OVERLOAD] Restart failed for ${name} (pm_id: ${pm_id}):`,
+            restartErr,
+          )
+
+          setCpuOverloadRestartFailedHistory(pm_id)
+        }
+      }
+    }
+  } catch (err) {
+    logger('error', err)
+  } finally {
+    isProcessCheckerRunning = false
+  }
+}
+
+const runModule = () => {
+  if (!MODULE_ENABLED) return
+
+  // connect to local pm2
+  pm2.connect((err) => {
+    if (err) {
+      logger('error', `PM2 connection error:`, err)
+
+      process.exit(1)
+    }
+
+    logger('info', 'Connected to PM2, starting monitor...')
+
+    processChecker()
+
+    setInterval(() => {
+      processChecker()
+    }, WORKER_INTERVAL)
+  })
+
+  /** PROB PMX **/
+  Probe.metric({
+    name: 'Zombie Restarts',
+    value: () => {
+      const res = []
+
+      for (const [k, v] of zombieRestartHistory) {
+        if (v > 0) {
+          res.push([k, v])
+        }
+      }
+
+      if (!res.length) return 'N/A'
+
+      return res.map((v) => `[${v[0]}]:${v[1]}`).join(';')
+    },
+  })
+
+  Probe.metric({
+    name: 'Zombie Restarts (failed)',
+    value: () => {
+      const res = []
+
+      for (const [k, v] of zombieRestartFailedHistory) {
+        if (v > 0) {
+          res.push([k, v])
+        }
+      }
+
+      if (!res.length) return 'N/A'
+
+      return res.map((v) => `[${v[0]}]:${v[1]}`).join(';')
+    },
+  })
+
+  Probe.metric({
+    name: 'CPU Overload Restarts',
+    value: () => {
+      const res = []
+
+      for (const [k, v] of cpuOverloadRestartHistory) {
+        if (v > 0) {
+          res.push([k, v])
+        }
+      }
+
+      if (!res.length) return 'N/A'
+
+      return res.map((v) => `[${v[0]}]:${v[1]}`).join(';')
+    },
+  })
+
+  Probe.metric({
+    name: 'CPU Overload Restarts (failed)',
+    value: () => {
+      const res = []
+
+      for (const [k, v] of cpuOverloadRestartFailedHistory) {
+        if (v > 0) {
+          res.push([k, v])
+        }
+      }
+
+      if (!res.length) return 'N/A'
+
+      return res.map((v) => `[${v[0]}]:${v[1]}`).join(';')
+    },
+  })
+
+  Probe.metric({
+    name: 'Processes in Sampling (perf)',
+    value: () => {
+      const res = []
+
+      for (const [k, v] of perfSamplingStats) {
+        if (v === true) {
+          res.push(k)
+        }
+      }
+
+      if (!res.length) return 'N/A'
+
+      return res.join(', ')
+    },
+  })
+
+  Probe.metric({
+    name: 'CPU (pm2)',
+    value: () => {
+      const res = []
+
+      for (const [k, v] of cpuUsageFromPM2) {
+        res.push([k, v])
+      }
+
+      if (!res.length) return 'N/A'
+
+      return res.map((v) => `[${v[0]}]:${v[1]}%`).join(';')
+    },
+  })
+
+  Probe.metric({
+    name: 'CPU (sys)',
+    value: () => {
+      const res = []
+
+      for (const [k, v] of cpuUsageFromSys) {
+        res.push([k, v])
+      }
+
+      if (!res.length) return 'N/A'
+
+      return res.map((v) => `[${v[0]}]:${v[1]}%`).join(';')
+    },
+  })
+}
+
+runModule()