Nginx主动健康检查的实战指南

Nginx 主动健康检查实战指南

主动健康检查是 Nginx Plus 的专有功能，开源版 Nginx 需通过第三方模块或搭配其他工具实现。以下是两种方案的详细指南：

一、Nginx Plus 原生方案

upstream backend {
    zone backend_servers 64k;

    server backend1.example.com:80 resolve;
    server backend2.example.com:80 resolve;

    # 主动健康检查配置
    health_check interval=5s 
                 passes=3 
                 fails=2 
                 uri=/health
                 match=status_ok;
}

# 健康检查匹配条件
match status_ok {
    status 200;
    body ~ "healthy";
    header Content-Type = text/html;
}

http {
    upstream myapp {
        zone myapp_zone 64k;
        least_conn;

        server 10.0.0.1:8080 slow_start=30s;
        server 10.0.0.2:8080 max_fails=3 fail_timeout=30s;
        server backup.example.com:8080 backup;

        # 主动健康检查
        health_check interval=10s
                     jitter=2s
                     fails=2
                     passes=1
                     uri=/api/health
                     port=8080
                     match=health_check;
    }

    match health_check {
        status 200-399;
        header Cache-Control ~ "no-cache";
        body !~ "maintenance";
    }

    server {
        listen 80;

        location / {
            proxy_pass http://myapp;
            proxy_set_header Host $host;
            proxy_next_upstream error timeout http_500;
        }

        # 健康状态页面（可选）
        location /upstream_status {
            status_zone upstream_status;
            proxy_pass http://myapp;
        }
    }
}

二、开源 Nginx 替代方案

# 编译安装
cd nginx-1.20.1
patch -p1 < /path/to/nginx_upstream_check_module/check_1.20.1+.patch
./configure --add-module=/path/to/nginx_upstream_check_module
make && make install

upstream backend {
    server 192.168.1.100:80;
    server 192.168.1.101:80;

    check interval=3000 
          rise=2 
          fall=3 
          timeout=1000 
          type=http;
    check_http_send "GET /health HTTP/1.0\r\n\r\n";
    check_http_expect_alive http_2xx http_3xx;
}

upstream backend {
    server 10.0.0.1:80 max_fails=1 fail_timeout=10s;
    server 10.0.0.2:80;

    hc interval=5s 
       timeout=1s 
       type=http 
       port=80 
       uri=/health 
       status=200 
       up_status=up 
       down_status=down;
}

三、实战场景配置

# 微服务专用匹配条件
match microservice_health {
    status 200;
    header Content-Type ~ "application/json";
    body ~ '"status":"UP"';
    body !~ '"outOfService"';
}

upstream account_service {
    zone account_zone 128k;

    server account-svc-1:8080;
    server account-svc-2:8080;
    server account-svc-3:8080;

    health_check interval=3s
                 uri=/actuator/health
                 match=microservice_health
                 fails=1
                 passes=2;
}

stream {
    upstream db_backend {
        zone db_zone 64k;
        server db1.example.com:3306;
        server db2.example.com:3306;

        health_check interval=30s
                     port=3306
                     passes=1
                     fails=2
                     match=mysql_check;
    }

    match mysql_check {
        send "\x00\x00\x00\x0a\x40\x00\x00\x00\x00\x00\x00\x00";
        expect ~ "MySQL";
    }

    server {
        listen 3306;
        proxy_pass db_backend;
    }
}

四、监控与告警

# Nginx Plus 状态 API
location /api/status {
    api write=on;
    allow 10.0.0.0/8;
    deny all;
}

location = /status.html {
    root /usr/share/nginx/html;
    status_format html;
}

# nginx-prometheus-exporter 配置
scrape_configs:
  - job_name: 'nginx-plus'
    static_configs:
      - targets: ['nginx-host:8080']
    metrics_path: /api/6/http/upstreams
    params:
      upstream: ['backend']

groups:
  - name: nginx_alerts
    rules:
      - alert: NginxUpstreamUnhealthy
        expr: nginxplus_upstream_peer_unavail > 0
        for: 2m
        labels:
          severity: critical
        annotations:
          summary: "{{ $labels.upstream }} upstream has unhealthy nodes"

五、最佳实践建议

检查端点设计

• 专用健康检查端点（如 /health）
• 避免检查主业务接口
• 包含依赖服务状态（DB、Redis等）

参数调优

# 生产环境推荐值
health_check interval=5s  # 不宜过短，避免压力
              fails=3     # 避免抖动误判
              passes=2    # 确保稳定恢复
              timeout=2s  # 根据业务调整

灰度切换策略

upstream backend {
    server new-version weight=10 slow_start=60s;
    server old-version weight=90;

    health_check uri=/health gradual_start=on;
}

故障处理策略

server {
    proxy_next_upstream error timeout http_502 http_503;
    proxy_next_upstream_timeout 2s;
    proxy_next_upstream_tries 3;
}

六、常见问题排查

检查不生效

# 验证配置
nginx -t

# 查看日志
tail -f /var/log/nginx/error.log | grep health_check

# 检查共享内存
nginx -V 2>&1 | grep zone

性能优化

• 调整 zone 大小：zone backend 1M;
• 合理设置检查间隔，避免频繁请求
• 使用 jitter 分散检查时间

测试命令

# 手动触发检查
curl http://nginx/api/3/http/upstreams/backend/peer/1/health

# 查看状态
curl http://nginx/status

七、集成方案对比

八、安全注意事项

访问控制

location /health {
    internal;  # 限制内部访问
    allow 10.0.0.0/8;
    deny all;
}

敏感信息保护

# 健康检查接口不应泄露敏感信息
location = /health {
    return 200 "OK";
    add_header Content-Type text/plain;
}

通过以上配置和实践，可以构建健壮的主动健康检查机制，确保服务的高可用性。根据实际环境选择合适的方案，并做好监控告警，形成完整的健康管理闭环。