秒杀优化+秒杀安全

1.Redis预减库存

1.OrderServiceImpl.java 问题分析

2.具体实现 SeckillController.java

1.实现InitializingBean接口的afterPropertiesSet方法，在bean初始化之后将库存信息加载到Redis

    /*** 系统初始化，将秒杀商品库存加载到redis中** @throws Exception*/@Overridepublic void afterPropertiesSet() throws Exception {// 将秒杀商品库存加载到redis中List<GoodsVo> goodsVoList = goodsService.findGoodsVo();// 如果没有秒杀商品，直接返回if (CollectionUtils.isEmpty(goodsVoList)) {return;}goodsVoList.forEach(goodsVo -> {redisTemplate.opsForValue().set("seckillGoods:" + goodsVo.getId(), goodsVo.getStockCount());});}

2.进行库存预减

        // 库存预减Long stock = redisTemplate.opsForValue().decrement("seckillGoods:" + goodsId);// 判断库存是否充足if (stock < 0) {// 库存不足，返回秒杀失败页面redisTemplate.opsForValue().increment("seckillGoods:" + goodsId);model.addAttribute("errmsg", RespBeanEnum.EMPTY_STOCK.getMessage());return "secKillFail";}

3.优化分析

• 正常情况下，每次都需要到数据库减少库存，来解决超卖问题
• 使用Redis进行库存预减，可以减少对数据库的操作，从而提升效率

4.测试

1.清空Redis

2.清空订单表和秒杀商品表，设置一号商品库存为10

3.将项目部署上线

4.UserUtil.java生成100个用户

5.发送5000次请求

1.线程组配置

2.cookie管理器

3.秒杀请求

4.QPS为307，从80提升到了307提升了283%

5.但是，出现了库存遗留问题

5.缓存遗留原因分析

2.内存标记优化高并发

1.问题分析

• 在未使用内存标记时，每次请求都需要对库存进行预减，来判断是否有库存，即使库存为0
• 所以采用内存标记的方式，当库存为0的时候，就不用进行库存预减

2.具体实现 SeckillController.java

1.首先定义一个标记是否有库存的map

2.在系统初始化时，初始化map

3.如果库存预减发现没有库存了，就设置内存标记

4.在库存预减前，判断内存标记，减少redis访问

3.测试

1.将项目上线

2.清空订单表和秒杀商品表，设置一号商品库存为10

3.清空Redis

4.发送5000次请求，QPS为330，从307提高到了330

3.消息队列实现异步秒杀

1.问题分析

2.思路分析

3.构建秒杀消息对象 SeckillMessage.java

package com.sxs.seckill.pojo;import lombok.AllArgsConstructor;
import lombok.Data;
import lombok.NoArgsConstructor;/*** Description: 秒杀消息** @Author sun* @Create 2024/5/13 14:15* @Version 1.0*/
@Data
@NoArgsConstructor
@AllArgsConstructor
public class SeckillMessage {private User user;private Long goodsId;
}

4.秒杀RabbitMQ配置

package com.sxs.seckill.config;import org.springframework.amqp.core.Binding;
import org.springframework.amqp.core.BindingBuilder;
import org.springframework.amqp.core.Queue;
import org.springframework.amqp.core.TopicExchange;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;/*** Description: 秒杀RabbitMQ配置** @Author sun* @Create 2024/5/13 14:23* @Version 1.0*/
@Configuration
public class RabbitMQSeckillConfig {// 定义一个消息队列和一个topic交换机的名字public static final String SECKILL_QUEUE = "seckillQueue";public static final String SECKILL_EXCHANGE = "seckillExchange";// 创建一个消息队列@Beanpublic Queue seckillQueue() {return new Queue(SECKILL_QUEUE, true);}// 创建一个topic交换机@Beanpublic TopicExchange seckillExchange() {return new TopicExchange(SECKILL_EXCHANGE);}// 将消息队列绑定到交换机@Beanpublic Binding binding() {// 绑定消息队列到交换机，并指定routingKey，表示只接收routingKey为seckill.#的消息return BindingBuilder.bind(seckillQueue()).to(seckillExchange()).with("seckill.#");}
}

5.生产者和消费者

1.生产者 MQSendMessage.java

package com.sxs.seckill.rabbitmq;import lombok.extern.slf4j.Slf4j;
import org.springframework.amqp.rabbit.core.RabbitTemplate;
import org.springframework.stereotype.Service;import javax.annotation.Resource;/*** Description: 消息队列发送消息** @Author sun* @Create 2024/5/13 15:14* @Version 1.0*/
@Service
@Slf4j
public class MQSendMessage {@Resourceprivate RabbitTemplate rabbitTemplate;// 发送秒杀消息public void sendSeckillMessage(String message) {log.info("发送消息：" + message);rabbitTemplate.convertAndSend("seckillExchange", "seckill.message", message);}
}

2.消费者，进行秒杀

1.引入hutool工具类

        <dependency><groupId>cn.hutool</groupId><artifactId>hutool-all</artifactId><version>5.3.3</version></dependency>

2. MQReceiverMessage.java

package com.sxs.seckill.rabbitmq;import cn.hutool.json.JSONUtil;
import com.sxs.seckill.pojo.SeckillMessage;
import com.sxs.seckill.pojo.User;
import com.sxs.seckill.service.GoodsService;
import com.sxs.seckill.service.OrderService;
import com.sxs.seckill.service.SeckillGoodsService;
import com.sxs.seckill.vo.GoodsVo;
import lombok.extern.slf4j.Slf4j;
import org.springframework.amqp.rabbit.annotation.RabbitListener;
import org.springframework.stereotype.Service;import javax.annotation.Resource;/*** Description: 消息队列接收消息** @Author sun* @Create 2024/5/13 15:17* @Version 1.0*/
@Service
@Slf4j
public class MQReceiverMessage {@Resourceprivate GoodsService goodsService;@Resourceprivate OrderService orderService;// 接收秒杀消息@RabbitListener(queues = "seckillQueue")public void receiveSeckillMessage(String message) {log.info("接收消息：" + message);// 此时的message是秒杀的消息，要将其转换为SeckillMessage对象SeckillMessage seckillMessage = JSONUtil.toBean(message, SeckillMessage.class);// 获取秒杀信息User user = seckillMessage.getUser();Long goodsId = seckillMessage.getGoodsId();// 根据商品id查询商品详情GoodsVo goodsVoByGoodsId = goodsService.findGoodsVoByGoodsId(goodsId);// 进行秒杀orderService.seckill(user, goodsVoByGoodsId);}
}

6.编写控制层

1.SeckillController.java

        // MQ实现异步秒杀// 封装秒杀信息SeckillMessage seckillMessage = new SeckillMessage(user, goodsId);// 使用hutool工具类将SeckillMessage对象转换为json字符串并发送mqSendMessage.sendSeckillMessage(JSONUtil.toJsonStr(seckillMessage));// 返回排队中页面model.addAttribute("errmsg", RespBeanEnum.QUEUE_ERROR.getMessage());return "secKillFail";

2.RespBeanEnum.java 新增响应枚举类

7.测试

1.将项目上线

2.清空订单表和秒杀商品表，设置一号商品库存为10

3.清空Redis

4.发送5000次请求，QPS为363

秒杀安全

1.秒杀接口隐藏

1.需求分析

2.思路分析

3.具体实现

1.RespBeanEnum.java 新增几个响应

2.OrderService.java 新增方法

    /*** 方法：生成秒杀路径* @param user* @param goodsId* @return*/String createPath(User user, Long goodsId);/*** 方法：校验秒杀路径* @param user* @param goodsId* @param path* @return*/boolean checkPath(User user, Long goodsId, String path);

3.OrderServiceImpl.java

    @Overridepublic String createPath(User user, Long goodsId) {// 对参数进行校验if (user == null || goodsId <= 0) {return null;}// 生成秒杀路径String path = MD5Util.md5(UUIDUtil.uuid() + "123456");// 保存到redis中，设置过期时间为60秒redisTemplate.opsForValue().set("seckillPath:" + user.getId() + ":" + goodsId, path, 60, TimeUnit.SECONDS);return path;}@Overridepublic boolean checkPath(User user, Long goodsId, String path) {// 对参数进行校验if (user == null || goodsId <= 0 || StringUtils.isBlank(path)) {return false;}// 从redis中获取秒杀路径String redisPath = (String) redisTemplate.opsForValue().get("seckillPath:" + user.getId() + ":" + goodsId);// 判断是否相等，并返回return path.equals(redisPath);}

4.SeckillController.java

    @RequestMapping("/{path}/doSeckill")public RespBean doSeckill(Model model, User user, Long goodsId, @PathVariable String path) {// 判断用户是否登录if (user == null) {return RespBean.error(RespBeanEnum.SESSION_ERROR);}// 校验pathboolean check = orderService.checkPath(user, goodsId, path);if (!check) {return RespBean.error(RespBeanEnum.REQUEST_ILLEGAL);}// 根据goodsId获取GoodsVoGoodsVo goodsVoByGoodsId = goodsService.findGoodsVoByGoodsId(goodsId);// 判断是否有库存if (goodsVoByGoodsId.getStockCount() < 1) {return RespBean.error(RespBeanEnum.EMPTY_STOCK);}// 从redis中判断是否复购if (redisTemplate.hasKey("order:" + user.getId() + ":" + goodsId)) {return RespBean.error(RespBeanEnum.REPEATE_ERROR);}// 首先判断内存标记if (inventoryTagging.get(goodsId)) {return RespBean.error(RespBeanEnum.EMPTY_STOCK);}// 库存预减Long stock = redisTemplate.opsForValue().decrement("seckillGoods:" + goodsId);// 判断库存是否充足if (stock < 0) {// 标记库存不足inventoryTagging.put(goodsId, true);// 库存不足，返回秒杀失败页面redisTemplate.opsForValue().increment("seckillGoods:" + goodsId);return RespBean.error(RespBeanEnum.EMPTY_STOCK);}// MQ实现异步秒杀// 封装秒杀信息SeckillMessage seckillMessage = new SeckillMessage(user, goodsId);// 使用hutool工具类将SeckillMessage对象转换为json字符串并发送mqSendMessage.sendSeckillMessage(JSONUtil.toJsonStr(seckillMessage));// 返回排队中return RespBean.success(RespBeanEnum.SEK_KILL_WAIT);}/*** 生成秒杀地址* @param user* @param goodsId* @return*/@ResponseBody@RequestMapping("/path")public RespBean getPath(User user, Long goodsId) {// 参数校验if (user == null || goodsId <= 0) {return RespBean.error(RespBeanEnum.REQUEST_ILLEGAL);}// 调用OrderService中的createPath方法生成秒杀地址String path = orderService.createPath(user, goodsId);return RespBean.success(path);}

5.goodsDetail.html

1.秒杀首先获取路径

2.解析环境变量，区分多环境

3.新增两个方法，使用隐藏秒杀接口的方式秒杀商品

4.测试

2.验证码防止脚本攻击

1.思路分析

2.具体实现

1.pom.xml 引入依赖

        <dependency><groupId>com.ramostear</groupId><artifactId>Happy-Captcha</artifactId><version>1.0.1</version></dependency>

2.SeckillController.java 编写方法生成验证码

    /*** 生成验证码* @param user* @param goodsId* @param request* @param response*/@RequestMapping("/captcha")public void happyCaptcha(User user, Long goodsId, HttpServletRequest request, HttpServletResponse response) {HappyCaptcha.require(request, response).style(CaptchaStyle.ANIM) //设置展现样式为动画.type(CaptchaType.NUMBER) //设置验证码内容为数字.length(6) //设置字符长度为 6.width(220) //设置动画宽度为 220.height(80) //设置动画高度为 80.font(Fonts.getInstance().zhFont()) //设置汉字的字体.build().finish(); //生成并输出验证码// 这个验证码的结果会存储在session中，可以通过request.getSession().getAttribute("happy-captcha")获取// 获取验证码的值，放入redis中String verifyCode = request.getSession().getAttribute("happy-captcha").toString();redisTemplate.opsForValue().set("captcha:" + user.getId() + ":" + goodsId, verifyCode, 60, TimeUnit.SECONDS);}

3.OrderService.java 校验用户输入的验证码

    /*** 校验用户输入的验证码* @param user* @param goodsId* @param captcha* @return*/boolean checkCaptcha(User user, Long goodsId, String captcha);

4.OrderServiceImpl.java

    @Overridepublic boolean checkCaptcha(User user, Long goodsId, String captcha) {// 参数校验if (user == null || goodsId <= 0 || StringUtils.isBlank(captcha)) {return false;}// 从redis中获取验证码String verifyCode = (String) redisTemplate.opsForValue().get("captcha:" + user.getId() + ":" + goodsId);return captcha.equals(verifyCode);}

5.SeckillController.java 加入验证码校验

6.goodsDetail.html

1.前端请求验证码

2.测试

3.获取用户输入的验证码，并携带验证码

3.秒杀接口限流-防刷

1.思路分析

2.简单接口限流

1.SeckillController.java

2.测试

4.通用接口限流防刷

1.思路分析

2.编写自定义限流注解 AccessLimit.java

package com.sxs.seckill.config;import java.lang.annotation.ElementType;
import java.lang.annotation.Retention;
import java.lang.annotation.RetentionPolicy;
import java.lang.annotation.Target;/*** Description: 限流注解** @Author sun* @Create 2024/5/14 15:38* @Version 1.0*/
@Retention(RetentionPolicy.RUNTIME) // 运行时生效
@Target(ElementType.METHOD) // 作用在方法上
public @interface AccessLimit {int seconds(); // 时间范围int maxCount(); // 最大访问次数boolean needLogin() default true; // 是否需要登录
}

3.使用方式 SeckillController.java

4.编写 config/UserContext.java 使用ThreadLocal存储user

package com.sxs.seckill.config;import com.sxs.seckill.pojo.User;/*** Description:** @Author sun* @Create 2024/5/14 15:46* @Version 1.0*/
public class UserContext {// 初始化ThreadLocal以存储用户信息private static ThreadLocal<User> threadLocal = new ThreadLocal<>();public static User getUser() {return threadLocal.get();}public static void setUser(User user) {threadLocal.set(user);}// 清除ThreadLocal中的数据public static void removeUser() {threadLocal.remove();}
}

5.编写自定义限流拦截器 config/AccessLimitInterceptor.java

package com.sxs.seckill.config;import com.sxs.seckill.exception.GlobalException;
import com.sxs.seckill.pojo.User;
import com.sxs.seckill.service.UserService;
import com.sxs.seckill.utils.CookieUtil;
import com.sxs.seckill.vo.RespBeanEnum;
import org.springframework.data.redis.core.RedisTemplate;
import org.springframework.stereotype.Component;
import org.springframework.web.method.HandlerMethod;
import org.springframework.web.servlet.HandlerInterceptor;import javax.annotation.Resource;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;
import java.util.concurrent.TimeUnit;/*** Description: 限流拦截器** @Author sun* @Create 2024/5/14 15:55* @Version 1.0*/
@Component
public class AccessLimitInterceptor implements HandlerInterceptor {@Resourceprivate UserService userService;@ResourceRedisTemplate redisTemplate;/*** 拦截请求，进行限流处理，在目标方法前执行** @param request* @param response* @param handler* @return* @throws Exception*/@Overridepublic boolean preHandle(HttpServletRequest request, HttpServletResponse response, Object handler) throws Exception {if (handler instanceof HandlerMethod) {// 如果是方法级别的拦截// 1.获取user对象，放到threadLocal中User user = getUser(request, response);UserContext.setUser(user);// 2.处理限流注解HandlerMethod handlerMethod = (HandlerMethod) handler;AccessLimit accessLimit = handlerMethod.getMethodAnnotation(AccessLimit.class);if (accessLimit == null) {return true;}// 3.获取注解上的参数int seconds = accessLimit.seconds();int maxCount = accessLimit.maxCount();boolean needLogin = accessLimit.needLogin();String key = request.getRequestURI();if (needLogin) {// 如果需要登录，但是没有登录，返回错误信息if (user == null) {// 如果需要登录，但是没有登录，返回错误信息throw new GlobalException(RespBeanEnum.USER_NOT_LOGIN);}// 如果登录了，key加上用户idkey += ":" + user.getId();}// 4.对访问次数进行限制，如果登陆了就是对这个用户的访问次数进行限制，如果没有登录就是对这个接口的访问次数进行限制Integer count = (Integer) redisTemplate.opsForValue().get(key);if (count == null) {// 第一次访问redisTemplate.opsForValue().set(key, 1, seconds, TimeUnit.SECONDS);} else if (count < maxCount) {// 访问次数加1redisTemplate.opsForValue().increment(key);} else {// 超过访问次数throw new GlobalException(RespBeanEnum.ACCESS_LIMIT_REACHED);}}// 如果不是方法级别的拦截，直接放行return true;}// 单独编写方法，获取User对象private User getUser(HttpServletRequest request, HttpServletResponse response) {String ticket = CookieUtil.getCookieValue(request, "userTicket");if (ticket == null) {return null;}return userService.getUserByCookie(ticket, request, response);}
}

6.config/WebConfig.java中注册拦截器

7.修改自定义参数解析器UserArgumentResolver.java，直接从ThreadLocal中获取User

8.测试

9.解决库存遗留问题，为每个用户id加锁即可