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>
This commit is contained in:
LoganandClaude committed
1 parent
092d0e4b9e
commit
471353e3d7
190 files changed
+16640
-3484
No files matched your search
@@ -0,0 +1,181 @@
|
||||
package controller
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"encoding/json"
|
||||
"fmt"
|
||||
"io"
|
||||
"net/http"
|
||||
"strings"
|
||||
"time"
|
||||
|
||||
"ezops/model"
|
||||
|
||||
"gorm.io/gorm"
|
||||
)
|
||||
|
||||
// dispatchCategoryAlert 按告警集合策略发送告警通知 // Logan.Li
|
||||
func dispatchCategoryAlert(db *gorm.DB, category string, ruleName, hostname, metric, severity, message string, value, threshold float64) {
|
||||
var policy model.AlertNotifyPolicy
|
||||
if err := db.Where("category = ? AND enabled = ?", category, true).First(&policy).Error; err != nil {
|
||||
fallbackNotifyChannels(db, ruleName, hostname, metric, severity, message, value, threshold)
|
||||
return
|
||||
}
|
||||
title := buildNotifyTitle(db, ruleName, hostname, metric, severity)
|
||||
content := buildAlertContent(ruleName, hostname, metric, severity, message, value, threshold)
|
||||
sent, _ := dispatchToPolicyContacts(db, &policy, title, content, severity, false)
|
||||
if sent == 0 {
|
||||
fallbackNotifyChannels(db, ruleName, hostname, metric, severity, message, value, threshold)
|
||||
}
|
||||
}
|
||||
|
||||
// dispatchCategoryRecovery 按告警集合策略发送恢复通知 // Logan.Li
|
||||
func dispatchCategoryRecovery(db *gorm.DB, category, ruleName, hostname, metric, message string) {
|
||||
var policy model.AlertNotifyPolicy
|
||||
if err := db.Where("category = ? AND enabled = ?", category, true).First(&policy).Error; err != nil {
|
||||
fallbackNotifyRecovery(db, ruleName, hostname, metric, message)
|
||||
return
|
||||
}
|
||||
title := fmt.Sprintf("%s — 已恢复正常", ruleName)
|
||||
content := fmt.Sprintf("**告警已恢复**\\n\\n目标：%s\\n指标：%s\\n详情：%s\\n恢复时间：%s",
|
||||
hostname, metric, message, time.Now().Format("2006-01-02 15:04:05"))
|
||||
sent, _ := dispatchToPolicyContacts(db, &policy, title, content, "info", true)
|
||||
if sent == 0 {
|
||||
fallbackNotifyRecovery(db, ruleName, hostname, metric, message)
|
||||
}
|
||||
}
|
||||
|
||||
// resolvePolicyContactIDs 展开策略中的联系组与直接联系人 // Logan.Li
|
||||
func resolvePolicyContactIDs(db *gorm.DB, policy *model.AlertNotifyPolicy) []uint {
|
||||
idSet := make(map[uint]struct{})
|
||||
|
||||
var groupIDs []uint
|
||||
if policy.ContactGroupIDs != "" && policy.ContactGroupIDs != "null" {
|
||||
_ = json.Unmarshal([]byte(policy.ContactGroupIDs), &groupIDs)
|
||||
}
|
||||
for _, gid := range groupIDs {
|
||||
var g model.AlertContactGroup
|
||||
if db.First(&g, gid).Error != nil {
|
||||
continue
|
||||
}
|
||||
var cids []uint
|
||||
if g.ContactIDs != "" && g.ContactIDs != "null" {
|
||||
_ = json.Unmarshal([]byte(g.ContactIDs), &cids)
|
||||
}
|
||||
for _, cid := range cids {
|
||||
idSet[cid] = struct{}{}
|
||||
}
|
||||
}
|
||||
|
||||
var directIDs []uint
|
||||
if policy.ContactIDs != "" && policy.ContactIDs != "null" {
|
||||
_ = json.Unmarshal([]byte(policy.ContactIDs), &directIDs)
|
||||
}
|
||||
for _, cid := range directIDs {
|
||||
idSet[cid] = struct{}{}
|
||||
}
|
||||
|
||||
ids := make([]uint, 0, len(idSet))
|
||||
for id := range idSet {
|
||||
ids = append(ids, id)
|
||||
}
|
||||
return ids
|
||||
}
|
||||
|
||||
// dispatchToPolicyContacts 向策略解析出的联系人发送通知 // Logan.Li
|
||||
func dispatchToPolicyContacts(db *gorm.DB, policy *model.AlertNotifyPolicy, title, content, severity string, isRecovery bool) (sent, skipped int) {
|
||||
if policy == nil || !policy.Enabled {
|
||||
return 0, 1
|
||||
}
|
||||
contactIDs := resolvePolicyContactIDs(db, policy)
|
||||
if len(contactIDs) == 0 {
|
||||
return 0, 1
|
||||
}
|
||||
|
||||
for _, cid := range contactIDs {
|
||||
var contact model.AlertContact
|
||||
if db.First(&contact, cid).Error != nil || contact.WebhookURL == "" {
|
||||
skipped++
|
||||
continue
|
||||
}
|
||||
var err error
|
||||
switch contact.ChannelType {
|
||||
case "feishu", "":
|
||||
if isRecovery {
|
||||
err = sendFeishuRecovery(db, feishuConfigFromURL(contact.WebhookURL), title, content)
|
||||
} else {
|
||||
err = sendFeishuAlert(db, feishuConfigFromURL(contact.WebhookURL), title, content, severity)
|
||||
}
|
||||
case "dingtalk", "wechat", "webhook":
|
||||
err = sendGenericWebhook(contact.WebhookURL, title, content)
|
||||
default:
|
||||
err = sendGenericWebhook(contact.WebhookURL, title, content)
|
||||
}
|
||||
if err != nil {
|
||||
fmt.Printf("[AlertNotify] 通知失败 (联系人:%s): %v\n", contact.Name, err)
|
||||
skipped++
|
||||
} else {
|
||||
sent++
|
||||
}
|
||||
}
|
||||
return sent, skipped
|
||||
}
|
||||
|
||||
// feishuConfigFromURL 联系人 webhook 转飞书配置 JSON // Logan.Li
|
||||
func feishuConfigFromURL(url string) string {
|
||||
b, _ := json.Marshal(model.FeishuWebhookConfig{URL: url})
|
||||
return string(b)
|
||||
}
|
||||
|
||||
// sendGenericWebhook 钉钉/企微/通用 Webhook 文本通知 // Logan.Li
|
||||
func sendGenericWebhook(webhookURL, title, content string) error {
|
||||
body := map[string]interface{}{
|
||||
"msgtype": "text",
|
||||
"text": map[string]string{
|
||||
"content": title + "\n" + strings.ReplaceAll(content, "\\n", "\n"),
|
||||
},
|
||||
}
|
||||
raw, _ := json.Marshal(body)
|
||||
req, err := http.NewRequest("POST", webhookURL, bytes.NewReader(raw))
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
req.Header.Set("Content-Type", "application/json")
|
||||
client := &http.Client{Timeout: 10 * time.Second}
|
||||
resp, err := client.Do(req)
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
if resp.StatusCode != http.StatusOK {
|
||||
b, _ := io.ReadAll(resp.Body)
|
||||
return fmt.Errorf("HTTP %d: %s", resp.StatusCode, string(b))
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
// fallbackNotifyChannels 无策略时回退到全局 NotifyChannel // Logan.Li
|
||||
func fallbackNotifyChannels(db *gorm.DB, ruleName, hostname, metric, severity, message string, value, threshold float64) {
|
||||
var channels []model.NotifyChannel
|
||||
db.Where("enabled = ? AND type = ?", true, "feishu").Find(&channels)
|
||||
title := buildNotifyTitle(db, ruleName, hostname, metric, severity)
|
||||
content := buildAlertContent(ruleName, hostname, metric, severity, message, value, threshold)
|
||||
for _, ch := range channels {
|
||||
if err := sendFeishuAlert(db, ch.Config, title, content, severity); err != nil {
|
||||
fmt.Printf("[AlertNotify] 飞书回退通知失败 (渠道:%s): %v\n", ch.Name, err)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func fallbackNotifyRecovery(db *gorm.DB, ruleName, hostname, metric, message string) {
|
||||
var channels []model.NotifyChannel
|
||||
db.Where("enabled = ? AND type = ?", true, "feishu").Find(&channels)
|
||||
title := fmt.Sprintf("%s — 已恢复正常", ruleName)
|
||||
content := fmt.Sprintf("**告警已恢复**\\n\\n主机：%s\\n指标：%s\\n详情：%s\\n恢复时间：%s",
|
||||
hostname, metric, message, time.Now().Format("2006-01-02 15:04:05"))
|
||||
for _, ch := range channels {
|
||||
if err := sendFeishuRecovery(db, ch.Config, title, content); err != nil {
|
||||
fmt.Printf("[AlertNotify] 飞书恢复回退失败 (渠道:%s): %v\n", ch.Name, err)
|
||||
}
|
||||
}
|
||||
}
|
||||
Reference in new issue
Block a user