修复工单observer-23906843495三个Bug：告警阈值/IP子域名误判/邮件次数写死

Agent-Logs-Url: https://github.com/qinfendebingshuo/guanghulab/sessions/fd18e928-cba0-469e-b2c5-ce77d8efc2b1

Co-authored-by: qinfendebingshuo <207279273+qinfendebingshuo@users.noreply.github.com>

.github/workflows/zhuyuan-deploy-observer.yml

@@ -114,6 +114,7 @@ jobs:
     outputs:
       repair_success: ${{ steps.repair.outputs.repair_success }}
       repair_attempt: ${{ steps.repair.outputs.repair_attempt }}
+      needs_human: ${{ steps.repair.outputs.needs_human }}
 
     steps:
       - name: '📥 检出代码'
@@ -212,7 +213,7 @@ jobs:
     if: >-
       always() &&
       needs.analyze-and-repair.result != 'skipped' &&
-      needs.analyze-and-repair.outputs.repair_success != 'true'
+      needs.analyze-and-repair.outputs.needs_human == 'true'
     runs-on: ubuntu-latest
 
     steps:
@@ -269,7 +270,8 @@ jobs:
         run: |
           if [ -n "$ZY_SMTP_USER" ] && [ -n "$ZY_SMTP_PASS" ]; then
             node scripts/staging-ops-agent.js alert \
-              --order-id "observer-${{ needs.collect-logs.outputs.run_id }}" || true
+              --order-id "observer-${{ needs.collect-logs.outputs.run_id }}" \
+              --attempt "${{ needs.analyze-and-repair.outputs.repair_attempt }}" || true
           else
             echo "ℹ️ SMTP未配置 · 跳过邮件告警 · 已创建GitHub Issue"
           fi

data/work-orders/active.json

@@ -116,7 +116,7 @@
     {
       "id": "WO-20260402-1457",
       "title": "Merge pull request #258 from qinfendebingshuo/copilot/update-homepage-three-column-layout",
-      "status": "failed",
+      "status": "human-intervened",
       "commit_sha": "b3f38e263f6f6aa6e03a516445cbea8b63bd9c09",
       "branch": "main",
       "created_by": "qinfendebingshuo",
@@ -141,6 +141,12 @@
           "status": "failed",
           "actor": "Agent",
           "message": "部署失败"
+        },
+        {
+          "timestamp": "2026-04-02T16:33:46.728Z",
+          "status": "human-intervened",
+          "actor": "铸渊",
+          "message": "铸渊人工干预 · 已定位根因：健康检查IP子域名误判Bug + 告警阈值Bug · 正在修复"
         }
       ],
       "deploy_logs": [
@@ -150,7 +156,7 @@
         }
       ],
       "created_at": "2026-04-02T14:57:46.729Z",
-      "updated_at": "2026-04-02T14:58:27.523Z"
+      "updated_at": "2026-04-02T16:33:46.728Z"
     }
   ]
 }

scripts/deputy-auto-repair.js

@@ -284,6 +284,7 @@ async function repair() {
     console.log(`❌ 已达最大修复次数(${MAX_REPAIR_ATTEMPTS}) · 放弃修复 · 通知人类`);
     setOutput('repair_success', 'false');
     setOutput('repair_attempt', String(attemptNumber - 1));
+    setOutput('needs_human', 'true');
     return;
   }
 
@@ -460,16 +461,18 @@ async function repair() {
   }
 
   // §7 设置输出
+  const needsHuman = !repairSuccess && attemptNumber >= MAX_REPAIR_ATTEMPTS;
   setOutput('repair_success', repairSuccess ? 'true' : 'false');
   setOutput('repair_attempt', String(attemptNumber));
+  setOutput('needs_human', needsHuman ? 'true' : 'false');
 
   console.log('═'.repeat(60));
   if (repairSuccess) {
     console.log(`✅ 修复成功 · 第${attemptNumber}次尝试`);
-  } else if (attemptNumber >= MAX_REPAIR_ATTEMPTS) {
+  } else if (needsHuman) {
     console.log(`❌ ${MAX_REPAIR_ATTEMPTS}次修复均失败 · 需要人工干预`);
   } else {
-    console.log(`❌ 第${attemptNumber}次修复失败 · 还有${MAX_REPAIR_ATTEMPTS - attemptNumber}次尝试机会`);
+    console.log(`❌ 第${attemptNumber}次修复失败 · 还有${MAX_REPAIR_ATTEMPTS - attemptNumber}次尝试机会 · 等待下次重试`);
   }
 }
 

scripts/staging-ops-agent.js

@@ -100,14 +100,22 @@ async function runHealthCheck(host) {
   console.log(`🔍 健康检查 · 目标: ${host}`);
   console.log('─'.repeat(50));
 
+  // 如果host是IP地址，带host_prefix的检查无效（IP不支持子域名）→ 降级为non-critical
+  const isIpAddress = /^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}$/.test(host);
+  if (isIpAddress) {
+    console.log('  ℹ️ 目标为IP地址 · 子域名检查降级为非关键');
+  }
+
   const results = [];
   let criticalFail = false;
 
   for (const endpoint of HEALTH_ENDPOINTS) {
     const checkHost = endpoint.host_prefix ? `${endpoint.host_prefix}.${host}` : host;
+    // IP地址不支持子域名前缀，降级为non-critical避免误报
+    const isCritical = endpoint.critical && !(isIpAddress && endpoint.host_prefix);
     const result = await httpCheck(checkHost, endpoint.path, endpoint.port);
 
-    const icon = result.ok ? '✅' : (endpoint.critical ? '❌' : '⚠️');
+    const icon = result.ok ? '✅' : (isCritical ? '❌' : '⚠️');
     console.log(`  ${icon} ${endpoint.name}: ${result.ok ? 'OK' : result.error || `HTTP ${result.status}`}`);
 
     results.push({
@@ -116,7 +124,7 @@ async function runHealthCheck(host) {
       checked_at: new Date().toISOString()
     });
 
-    if (!result.ok && endpoint.critical) {
+    if (!result.ok && isCritical) {
       criticalFail = true;
     }
   }
@@ -302,16 +310,17 @@ async function analyzeLog(logContent, orderId) {
   return deepResult;
 }
 
-// ── 邮件告警 (3次重试失败) ──────────────────
+// ── 邮件告警 (最大重试失败) ──────────────────
-async function sendAlert(orderId, email) {
+async function sendAlert(orderId, email, options = {}) {
   const smtpUser = process.env.ZY_SMTP_USER;
   const smtpPass = process.env.ZY_SMTP_PASS;
+  const attemptCount = options.attempt || process.env.REPAIR_ATTEMPT || '?';
 
   if (!smtpUser || !smtpPass) {
     console.error('❌ SMTP未配置 (需要ZY_SMTP_USER和ZY_SMTP_PASS)');
     console.log('📧 告警内容 (控制台输出):');
     console.log(`   工单: ${orderId}`);
-    console.log(`   状态: 3次自动修复未能解决，需要人工干预`);
+    console.log(`   状态: ${attemptCount}次自动修复未能解决，需要人工干预`);
     return false;
   }
 
@@ -345,7 +354,7 @@ async function sendAlert(orderId, email) {
     <p style="color: #666; margin-top: 0;">ZY-Staging-Ops-Agent · 自动告警</p>
 
     <div style="background: #fef5f5; border: 1px solid #f5c6cb; border-radius: 8px; padding: 15px; margin: 20px 0;">
-      <strong style="color: #c0392b;">⚠️ 自动修复已达最大重试次数(3次)，需要冰朔人工干预</strong>
+      <strong style="color: #c0392b;">⚠️ 自动修复已达最大重试次数(${attemptCount}次)，需要冰朔人工干预</strong>
     </div>
 
     <table style="width: 100%; border-collapse: collapse;">
@@ -473,7 +482,7 @@ async function main() {
         console.error('❌ 需要 --order-id 参数');
         process.exit(1);
       }
-      await sendAlert(orderId, args.email);
+      await sendAlert(orderId, args.email, { attempt: args.attempt });
       break;
     }