SpringBoot基于RateLimiter+AOP动态的为不同接口限流

水深无声 2021-12-07 11:17 205阅读 0赞

公司需要对接第三方接口,每个接口都有频率限制,只好做下精准的限流,精准到方法,如果是服务的话Hystrix应该可以的,废话不多说,看了一篇不错的文章,亲测可行,记录一下.  
1.首先接口限流算法：

1.计数器方式（传统计数器缺点：临界问题 可能违背定义固定速率原则）

![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L20wXzM3OTcwNjk5_size_16_color_FFFFFF_t_70]  
2.令牌桶方式  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L20wXzM3OTcwNjk5_size_16_color_FFFFFF_t_70 1]

3.漏桶方式

![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L20wXzM3OTcwNjk5_size_16_color_FFFFFF_t_70 2]  
4.应用层限流（Nginx）  
2.限流实现：

2.1. RateLimiter是guava提供的基于令牌桶算法的实现类，可以非常简单的完成限流特技，并且根据系统的实际情况来调整生成token的速率。
    
    2.2.导入相关依赖包
    
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-web</artifactId>
    </dependency>
     
    <dependency>
        <groupId>org.aspectj</groupId>
        <artifactId>aspectjweaver</artifactId>
    </dependency>
     
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-test</artifactId>
        <scope>test</scope>
    </dependency>
     
    <dependency>
        <groupId>com.google.guava</groupId>
        <artifactId>guava</artifactId>
        <version>20.0</version>
    </dependency>

2.3.代码实现不多说每一步都有注解

2.3.1 定义注解
    
        @Inherited
        @Documented
        @Target(ElementType.METHOD)
        @Retention(RetentionPolicy.RUNTIME)
        public @interface RateLimit {
            double limitNum() default 20;  //默认每秒放入桶中的token
        }
    
    2.3.2 封装定义返回结果
    
        public class MyResult {
            private Integer status;
            private String msg;
            private List<Object> data;
     
            public MyResult(Integer status, String msg, List<Object> data) {
                this.status = status;
                this.msg = msg;
                this.data = data;
            }
     
            public static MyResult OK(String msg, List<Object> data) {
                return new MyResult(200, msg, data);
            }
     
            public static MyResult Error(Integer status, String msg) {
                return new MyResult(status, msg, null);
            }

2.3.3 aop实现

@Component
        @Scope
        @Aspect
        public class RateLimitAspect {
            private Logger log = LoggerFactory.getLogger(this.getClass());
            //用来存放不同接口的RateLimiter(key为接口名称，value为RateLimiter)
            private ConcurrentHashMap<String, RateLimiter> map = new ConcurrentHashMap<>();
         
            private static ObjectMapper objectMapper = new ObjectMapper();
         
            private RateLimiter rateLimiter;
         
            @Autowired
            private HttpServletResponse response;
         
            @Pointcut("@annotation(com.icat.retalimitaop.annotation.RateLimit)")
            public void serviceLimit() {
            }
         
            @Around("serviceLimit()")
            public Object around(ProceedingJoinPoint joinPoint) throws NoSuchMethodException {
                Object obj = null;
                //获取拦截的方法名
                Signature sig = joinPoint.getSignature();
                //获取拦截的方法名
                MethodSignature msig = (MethodSignature) sig;
                //返回被织入增加处理目标对象
                Object target = joinPoint.getTarget();
                //为了获取注解信息
                Method currentMethod = target.getClass().getMethod(msig.getName(), msig.getParameterTypes());
                //获取注解信息
                RateLimit annotation = currentMethod.getAnnotation(RateLimit.class);
                double limitNum = annotation.limitNum(); //获取注解每秒加入桶中的token
                String functionName = msig.getName(); // 注解所在方法名区分不同的限流策略
         
                //获取rateLimiter
                 if(map.containsKey(functionName)){
                     rateLimiter = map.get(functionName);
                 }else {
                     map.put(functionName, RateLimiter.create(limitNum));
                     rateLimiter = map.get(functionName);
                 }
         
                try {
                    if (rateLimiter.tryAcquire()) {
                        //执行方法
                        obj = joinPoint.proceed();
                    } else {
                        //拒绝了请求（服务降级）
                        String result = objectMapper.writeValueAsString(MyResult.Error(500, "系统繁忙！"));
                        log.info("拒绝了请求：" + result);
                        outErrorResult(result);
                    }
                } catch (Throwable throwable) {
                    throwable.printStackTrace();
                }
                return obj;
            }
            //将结果返回
            public void outErrorResult(String result) {
                response.setContentType("application/json;charset=UTF-8");
                try (ServletOutputStream outputStream = response.getOutputStream()) {
                    outputStream.write(result.getBytes("utf-8"));
                } catch (IOException e) {
                    e.printStackTrace();
                }
            }
         
            static {
                objectMapper.setSerializationInclusion(JsonInclude.Include.NON_NULL);
            }
         
        }

3.测试限流

2个接口设定没秒限流5个和美妙限流10个 
        @RateLimit(limitNum = 5.0)
        public MyResult getResults() {
            log.info("调用了方法getResults");
            return MyResult.OK("调用了方法", null);
        }
     
        @RateLimit(limitNum = 10.0)
        public MyResult getResultTwo() {
            log.info("调用了方法getResultTwo");
            return MyResult.OK("调用了方法getResultTwo", null);
       }

使用postman测试getResults接口 20个并发（设定每秒只能处理5个请求）  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L20wXzM3OTcwNjk5_size_16_color_FFFFFF_t_70 3]  
使用postman测试getResultTwo接口 20个并发（设定每秒只能处理10个请求）  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L20wXzM3OTcwNjk5_size_16_color_FFFFFF_t_70 4]转自:[https://blog.csdn.net/qq\_39816039/article/details/83988517][https_blog.csdn.net_qq_39816039_article_details_83988517]

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L20wXzM3OTcwNjk5_size_16_color_FFFFFF_t_70]: /images/20211207/bb3703fd92d043058105bda2147fd2eb.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L20wXzM3OTcwNjk5_size_16_color_FFFFFF_t_70 1]: /images/20211207/d97a71501e304c0587e926f2af64c1db.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L20wXzM3OTcwNjk5_size_16_color_FFFFFF_t_70 2]: /images/20211207/8c72c23173fb4aeb89756d39d66c9f22.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L20wXzM3OTcwNjk5_size_16_color_FFFFFF_t_70 3]: /images/20211207/8d8ca1606fde4a17a798a790fcd09f4c.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L20wXzM3OTcwNjk5_size_16_color_FFFFFF_t_70 4]: https://img-blog.csdnimg.cn/20190716190145422.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L20wXzM3OTcwNjk5,size_16,color_FFFFFF,t_70
[https_blog.csdn.net_qq_39816039_article_details_83988517]: /images/20211207/c8d1f4c9c3b24c358bad51cdb3b6c658.png