后端某些接口在高并发的压力下往往会导致性能的严重下降,为了维持我们后端服务型的高性能和高可用,我们往往可以对某些接口或某些用户去设计限流机制,控制这些热点接口的访问量,我这里利用Redis的高性能优势,并整合AOP编程和引入Lua限流脚本在SpringBoot中对任意接口或某些用户实现了访问量限流的机制,其中,我这里给出了三种限流机制:用户,IP地址,全局限流
1.定义限流方式
/** * 限流类型 * @Author GuihaoLv */ public enum LimitType { /** * 默认策略全局限流 */ DEFAULT, /** * 根据请求者IP进行限流 */ IP, /** * 根据请求者的用户ID进行限流 */ USER, /** * 根据请求者的部门进行限流 */ DEPT, }
2.引入AOP,自定义限流注解和限流处理的切面类
/** * 限流注解 * @Author GuihaoLv */ //实例 @RateLimiter(time = 60, count = 5, limitType = LimitType.IP) 效果:同一IP 60秒内最多允许5次登录尝试。 @Target(ElementType.METHOD) //表示该注解仅能标注在方法上,用于对具体方法进行限流控制。 @Retention(RetentionPolicy.RUNTIME) //注解在运行时保留,可通过反射机制读取注解信息,实现动态限流逻辑。 @Documented //注解信息会包含在生成的 JavaDoc 中 public @interface RateLimiter { /** * 限流key */ public String key() default RedisConstant.RATE_LIMIT_KEY; /** * 限流时间,单位秒 */ public int time() default 60; /** * 限流次数 */ public int count() default 100; /** * 限流类型 */ public LimitType limitType() default LimitType.DEFAULT; }
/** * 限流处理切面 * @Author GuihaoLv */ @Aspect @Component //确保仅当配置项 spring.cache.type=redis 时,切面才会生效 @ConditionalOnProperty(prefix = "spring.cache", name = { "type" }, havingValue = "redis", matchIfMissing = false) public class RateLimiterAspect { private static final Logger log = LoggerFactory.getLogger(RateLimiterAspect.class); private RedisTemplate<Object, Object> redisTemplate; //redis 操作模板,用于执行 Lua 脚本和 Redis 命令 private RedisScript<Long> limitScript; //限流核心逻辑的 Lua 脚本 @Autowired public void setRedisTemplate1(RedisTemplate<Object, Object> redisTemplate) { this.redisTemplate = redisTemplate; } @Autowired public void setLimitScript(RedisScript<Long> limitScript) { this.limitScript = limitScript; } @Autowired private ObjectMapper jacksonObjectMapper; @Before("@annotation(rateLimiter)") public void doBefore(JoinPoint point, RateLimiter rateLimiter) throws Throwable { //1. 获取注解参数 int time = rateLimiter.time(); //时间窗口(秒) int count = rateLimiter.count();// 允许的请求次数 //2. 生成唯一限流 Key String combineKey = getCombineKey(rateLimiter, point); List<Object> keys = Collections.singletonList(combineKey); try { //3. 以key为参数执行 Lua 脚本(原子性操作) keys:Redis 存储的 Key //Lua脚本会检查Key是否存在,如果不存在则创建并设置过期时间,如果存在则递增计数器。 Long number = redisTemplate.execute(limitScript, keys, count, time); if (StringUtils.isEmpty(number) || number.intValue() > count) { throw new Exception("访问过于频繁,请稍候再试"); } log.info("限制请求'{}',当前请求'{}',缓存key'{}'", count, number.intValue(), combineKey); } catch (RuntimeException e) { throw new RuntimeException("服务器限流异常,请稍候再试"); } catch (Exception e) { throw e; } } //IP + 类名 + 方法名 的拼接方式,确保不同场景的Key不冲突。 //Key结构清晰,便于调试和监控(如通过Redis直接查看计数器)。 public String getCombineKey(RateLimiter rateLimiter, JoinPoint point) { StringBuffer stringBuffer = new StringBuffer(rateLimiter.key()); switch (rateLimiter.limitType()) { case IP: try { limitByIp(stringBuffer); } catch (IOException e) { e.printStackTrace(); } catch (InterruptedException e) { e.printStackTrace(); } break; case USER: limitByUser(stringBuffer); break; case DEFAULT: limitByDefault(stringBuffer,point); break; } return stringBuffer.toString(); } /** * 按IP限流 * @param stringBuffer */ private final HttpClient httpClient = HttpClient.newHttpClient(); private void limitByIp(StringBuffer stringBuffer) throws IOException, InterruptedException { String[] services = { "https://api.ipify.org", "https://icanhazip.com" }; for (String serviceUrl : services) { HttpRequest request = HttpRequest.newBuilder() .uri(URI.create(serviceUrl)) .GET() .build(); HttpResponse<String> response = httpClient.send(request, HttpResponse.BodyHandlers.ofString()); String ip = response.body().trim(); stringBuffer.append(":ip:").append(ip); } } /** * 全局限流 * @param stringBuffer * @param point */ private void limitByDefault(StringBuffer stringBuffer, JoinPoint point) { //按方法限流:拼接类名 + 方法名 MethodSignature signature = (MethodSignature) point.getSignature(); Method method = signature.getMethod(); Class<?> targetClass = method.getDeclaringClass(); stringBuffer.append(targetClass.getName()).append("-").append(method.getName()); } /** * 按用户限流 * @param stringBuffer */ private void limitByUser(StringBuffer stringBuffer) { //获取当前用户 String userSubject = UserThreadLocal.getSubject(); User user=new User(); try { user = jacksonObjectMapper.readValue(userSubject, User.class); } catch (JsonProcessingException e) { throw new RuntimeException("无法获取当前用户"); } stringBuffer.append(":user:").append(user.getId()); } }
3. 在Redis的配置类中整合Lua语言自定义限流脚本的执行
/** * 限流脚本定义 * @return */ @Bean public DefaultRedisScript<Long> limitScript() { DefaultRedisScript<Long> redisScript = new DefaultRedisScript<>(); redisScript.setScriptText(limitScriptText()); //加载Lua脚本 redisScript.setResultType(Long.class); // 返回类型为Long return redisScript; } /** * 限流脚本 */ private String limitScriptText() { return "local key = KEYS[1]\n" + "local count = tonumber(ARGV[1])\n" + "local time = tonumber(ARGV[2])\n" + "local current = redis.call('get', key);\n" + "if current and tonumber(current) > count then\n" + " return tonumber(current);\n" + "end\n" + "current = redis.call('incr', key)\n" + "if tonumber(current) == 1 then\n" + " redis.call('expire', key, time)\n" + "end\n" + "return tonumber(current);"; }
4. 限流注解的使用
案例1:登录接口IP限流
@RateLimiter( key = "login_attempt", time = 300, // 5分钟 count = 5, limitType = LimitType.IP ) @PostMapping("/login") public Response login(@RequestBody LoginDTO dto) { // 登录逻辑 }
案例2:API用户维度限流
@RateLimiter( key = "api_v1:data_export", time = 3600, // 1小时 count = 10,
limitType = LimitType.USER )
@GetMapping("/export")
public void exportData() { // 数据导出逻辑 }
这样就能对上述接口实现对应的限流机制了