Files
aiops/backend/pkg/siteprobe/worker.go
T
LoganandClaude 471353e3d7 feat: 监控告警体系统一重构 + 自动化任务调度引擎 + 监控大屏拆分
告警中心
- 新增统一告警(alert_unified)、监控告警规则(monitor_alert_rule)、指标目录(alert_metric_catalog)
- 通知策略与消息分发(alert_notify_policy/dispatch/message)、静默规则
- 告警引擎按阈值/SNMP 拆分(alert_engine_monitor/snmp)

监控采集
- SNMP 常量/模板与历史指标(snmp templates/constants、snmp_metrics_history、snmp_collect)
- 中间件监控 Redis 采集与种子(middleware_monitor_redis/seed)
- 站点探测 pkg(siteprobe) + 站点视图(site_view)

自动化任务
- cron 调度器(cronscheduler)、脚本运行器(scriptrunner)、系统任务(systask)及运行记录(cron_job_run、system_task)

审计与凭证
- 审计设置(audit_setting) + 历史清理(auditcleaner)

前端
- 监控页面(host/middleware/site/snmp)拆分为子模块并大屏化
- 告警中心多 Tab(规则/指标目录/通知策略/静默)、任务管理组件重构、Agent CLI Markdown 渲染
- 移除旧脚本卡片/详情/日志组件,统一为 ScriptTable/ScriptCodeEditor/TerminalOutput

Co-Authored-By: Claude <noreply@anthropic.com>
2026-07-01 23:44:45 +08:00

118 lines
2.6 KiB
Go

package siteprobe
import (
"context"
"log"
"sync"
"time"
"ezops/model"
"gorm.io/gorm"
)
// Worker 站点拨测后台采集器 // Logan.Li
type Worker struct {
db *gorm.DB
stopCh chan struct{}
wg sync.WaitGroup
}
// NewWorker 创建拨测 worker
func NewWorker(db *gorm.DB) *Worker {
return &Worker{db: db, stopCh: make(chan struct{})}
}
// Start 启动拨测循环
func (w *Worker) Start(intervalSec int) {
if intervalSec < 5 {
intervalSec = 10
}
w.wg.Add(1)
go func() {
defer w.wg.Done()
ticker := time.NewTicker(time.Duration(intervalSec) * time.Second)
defer ticker.Stop()
log.Printf("[SiteProbe] 拨测引擎启动,扫描间隔 %ds", intervalSec)
w.runOnce()
for {
select {
case <-ticker.C:
w.runOnce()
case <-w.stopCh:
log.Printf("[SiteProbe] 拨测引擎停止")
return
}
}
}()
}
// Stop 停止拨测
func (w *Worker) Stop() {
close(w.stopCh)
w.wg.Wait()
}
// RunTaskNow 立即执行单个任务拨测 // Logan.Li
func (w *Worker) RunTaskNow(taskID uint) error {
var task model.SiteTask
if err := w.db.First(&task, taskID).Error; err != nil {
return err
}
w.probeTask(task)
return nil
}
func (w *Worker) runOnce() {
var tasks []model.SiteTask
w.db.Where("status = ?", model.SiteTaskStatusRunning).Find(&tasks)
now := time.Now()
for _, task := range tasks {
if task.Frequency <= 0 {
task.Frequency = 60
}
if task.LastCheckAt != nil && now.Sub(*task.LastCheckAt) < time.Duration(task.Frequency)*time.Second {
continue
}
w.probeTask(task)
}
}
func (w *Worker) probeTask(task model.SiteTask) {
if task.TaskType == "" {
task.TaskType = model.SiteTaskType(inferTaskType(task))
}
outcome := Execute(context.Background(), task)
now := time.Now()
result := model.SiteResult{
TaskID: task.ID,
Status: outcome.Status,
HTTPStatusCode: outcome.HTTPStatus,
ResponseTime: outcome.ResponseTime,
TotalTime: outcome.TotalTime,
RemoteIP: outcome.RemoteIP,
ErrorMessage: outcome.ErrorMessage,
SSLCertExpiry: outcome.SSLCertExpiry,
SSLCertIssuer: outcome.SSLCertIssuer,
SSLCertSubject: outcome.SSLCertSubject,
SSLDaysRemaining: outcome.SSLDaysRemaining,
CheckedAt: now,
}
w.db.Create(&result)
lastResult := "fail"
if outcome.Status == model.SiteResultSuccess {
lastResult = "success"
}
avg := outcome.ResponseTime
if task.AvgResponse > 0 {
avg = (task.AvgResponse + outcome.ResponseTime) / 2
}
w.db.Model(&model.SiteTask{}).Where("id = ?", task.ID).Updates(map[string]interface{}{
"task_type": task.TaskType,
"last_result": lastResult,
"last_check_at": now,
"avg_response": avg,
})
}