SpringCloud gateway 限流源码分析

路由过滤器允许以某种方式修改传入的HTTP请求或传出的HTTP响应，路径过滤器的范围限定为特定路径，Spring Cloud Gateway包含许多内置的GatewayFilter工厂。

Spring Cloud Gateway限流就是通过内置的RequestRateLimiterGateWayFilterFactory工厂来实现的。

当然，官方的肯定不能满足我们部分业务需求，因此可以自定义限流过滤器。

## yml如下配置，就可以为该路由添加此拦截器：

spring:
   cloud:
      gateway:
        routes:
          - id: test_route
            uri: localhost
            predicates:
            - Path=/host/address
            filters:
            - name: RequestRateLimiter
              args:
                ## 允许用户每秒执行多少请求，而不会丢弃任何请求。这是令牌桶填充的速率。
                redis-rate-limiter.replenishRate: 1
                ## 是一秒钟内允许用户执行的最大请求数。这是令牌桶可以容纳的令牌数。将此值设置为零将阻止所有请求。
                redis-rate-limiter.burstCapacity: 3
                ## KeyResolver是一个简单的获取用户请求参数 我这里以主机地址为key来作限流
                key-resolver: "#{@hostAddrKeyResolver}"

## RequestRateLimiterGateWayFilterFactory代码：

//AbstractGatewayFilterFactory实现GatewayFilterFactory接口，自定义的过滤工厂可以继承
//AbstractGatewayFilterFactory并编写apply方法
public class RequestRateLimiterGatewayFilterFactory extends AbstractGatewayFilterFactory<RequestRateLimiterGatewayFilterFactory.Config> {

	public static final String KEY_RESOLVER_KEY = "keyResolver";

	private final RateLimiter defaultRateLimiter;
	private final KeyResolver defaultKeyResolver;

	public RequestRateLimiterGatewayFilterFactory(RateLimiter defaultRateLimiter,
												  KeyResolver defaultKeyResolver) {
		super(Config.class);
		this.defaultRateLimiter = defaultRateLimiter;
		this.defaultKeyResolver = defaultKeyResolver;
	}

	public KeyResolver getDefaultKeyResolver() {
		return defaultKeyResolver;
	}

	public RateLimiter getDefaultRateLimiter() {
		return defaultRateLimiter;
	}

	@SuppressWarnings("unchecked")
	@Override
	public GatewayFilter apply(Config config) {
                //yml中我们配置的hostAddrKeyResolver
		KeyResolver resolver = (config.keyResolver == null) ? defaultKeyResolver : config.keyResolver;
                //这个就是限流的具体实现，默认使用RedisRateLimiter
		RateLimiter<Object> limiter = (config.rateLimiter == null) ? defaultRateLimiter : config.rateLimiter;

		return (exchange, chain) -> {
			Route route = exchange.getAttribute(ServerWebExchangeUtils.GATEWAY_ROUTE_ATTR);

			return resolver.resolve(exchange).flatMap(key ->
                                    //这里的isAllowed就是具体实现，输入参数为路由id和限流key（这里为主机地址hostAddress）
					// TODO: if key is empty?
					limiter.isAllowed(route.getId(), key).flatMap(response -> {

						for (Map.Entry<String, String> header : response.getHeaders().entrySet()) {
							exchange.getResponse().getHeaders().add(header.getKey(), header.getValue());
						}
                                        //如果为真，通过拦截
						if (response.isAllowed()) {
							return chain.filter(exchange);
						}
                                        //否则设置http码为429，too many request
						exchange.getResponse().setStatusCode(config.getStatusCode());
						return exchange.getResponse().setComplete();
					}));
		};
	}


}

分析：

1.加载KeyResolver，从配置文件中加载，此处我配置了hostAddrKeyResolver，即根据host地址来进行限流。如果为空，使用默认的PrincipalNameKeyResolver

2.加载RateLimiter，默认使用RedisRateLimiter。

3.执行RedisRateLimiter的isAllowed方法，得到response，如果isAllowed为true则通过拦截，否则返回429（isAllowed方法具体实现下文描述）。

 

## HostAddrKeyResolver：

@Slf4j
public class HostAddrKeyResolver implements KeyResolver {
    @Override
    public Mono<String> resolve(ServerWebExchange exchange) {
        log.info("HostAddrKeyResolver 限流");
        return Mono.just(exchange.getRequest().getRemoteAddress().getHostName());
    }
}

在启动类中注入bean

@Bean
    public HostAddrKeyResolver hostAddrKeyResolver() {
        return new HostAddrKeyResolver();
    }

## RedisRateLimiter：

    @Override
	@SuppressWarnings("unchecked")
	public Mono<Response> isAllowed(String routeId, String id) {
                //判断是否初始化
		if (!this.initialized.get()) {
			throw new IllegalStateException("RedisRateLimiter is not initialized");
		}
                //获取配置
		Config routeConfig = getConfig().getOrDefault(routeId, defaultConfig);

		if (routeConfig == null) {
			throw new IllegalArgumentException("No Configuration found for route " + routeId);
		}

                //令牌桶填充速率
		int replenishRate = routeConfig.getReplenishRate();

                //令牌桶可容纳令牌数
		int burstCapacity = routeConfig.getBurstCapacity();

		try {
                    //获取redis的key，执行lua脚本时传入
			List<String> keys = getKeys(id);


                    //获取参数，执行lua脚本时传入
			List<String> scriptArgs = Arrays.asList(replenishRate + "", burstCapacity + "",
					Instant.now().getEpochSecond() + "", "1");
			Flux<List<Long>> flux = this.redisTemplate.execute(this.script, keys, scriptArgs);
					// .log("redisratelimiter", Level.FINER);
			return flux.onErrorResume(throwable -> Flux.just(Arrays.asList(1L, -1L)))
					.reduce(new ArrayList<Long>(), (longs, l) -> {
						longs.addAll(l);
						return longs;
					}) .map(results -> {
						boolean allowed = results.get(0) == 1L;
						Long tokensLeft = results.get(1);

						Response response = new Response(allowed, getHeaders(routeConfig, tokensLeft));

						if (log.isDebugEnabled()) {
							log.debug("response: " + response);
						}
						return response;
					});
		}
		catch (Exception e) {
			/*
			 * We don't want a hard dependency on Redis to allow traffic. Make sure to set
			 * an alert so you know if this is happening too much. Stripe's observed
			 * failure rate is 0.01%.
			 */
			log.error("Error determining if user allowed from redis", e);
		}
		return Mono.just(new Response(true, getHeaders(routeConfig, -1L)));
	}

	@NotNull
	public HashMap<String, String> getHeaders(Config config, Long tokensLeft) {
		HashMap<String, String> headers = new HashMap<>();
		headers.put(this.remainingHeader, tokensLeft.toString());
		headers.put(this.replenishRateHeader, String.valueOf(config.getReplenishRate()));
		headers.put(this.burstCapacityHeader, String.valueOf(config.getBurstCapacity()));
		return headers;
	}

	static List<String> getKeys(String id) {
		// use `{}` around keys to use Redis Key hash tags
		// this allows for using redis cluster

		// Make a unique key per user.
		String prefix = "request_rate_limiter.{" + id;
                //令牌桶剩余令牌数
		String tokenKey = prefix + "}.tokens";
                //令牌桶最后填充令牌时间
		String timestampKey = prefix + "}.timestamp";
		return Arrays.asList(tokenKey, timestampKey);
	}

分析：

1.判断是否初始化，加载配置，获取令牌填充速率和令牌桶大小

2.根据路由id组合成两个redis中的key值，传入lua脚本

request_rate_limiter.{id}.tokens 令牌桶剩余令牌数

request_rate_limiter.{id}.timestamp 令牌桶最后填充令牌时间

3.把令牌填充速率，令牌桶大小，当前时间（单位：秒），消耗令牌数（默认为1）组合传入lua脚本

4.执行lua脚本

5.flux.onErrorResume(throwable -> Flux.just(Arrays.asList(1L, -1L))) 这个是对执行lua脚本过程中发生异常的处理，它会忽略异常，返回令牌。这样就能跟redis解耦，不对它强依赖。

该实现核心主要体现在lua脚本上，它使用的是令牌桶算法

详见spring-cloud-gateway-core下的request_rate_limiter.lua

## 获取剩余令牌数的redis key
local tokens_key = KEYS[1]
## 获取最后一次填充令牌的时间
local timestamp_key = KEYS[2]
--redis.log(redis.LOG_WARNING, "tokens_key " .. tokens_key)

## 令牌填充速率
local rate = tonumber(ARGV[1])
## 令牌桶大小
local capacity = tonumber(ARGV[2])
## 当前秒数
local now = tonumber(ARGV[3])
## 消耗令牌数，默认1
local requested = tonumber(ARGV[4])

## 计算令牌桶需要填充的时间
local fill_time = capacity/rate
## 计算key的存活时间
local ttl = math.floor(fill_time*2)

--redis.log(redis.LOG_WARNING, "rate " .. ARGV[1])
--redis.log(redis.LOG_WARNING, "capacity " .. ARGV[2])
--redis.log(redis.LOG_WARNING, "now " .. ARGV[3])
--redis.log(redis.LOG_WARNING, "requested " .. ARGV[4])
--redis.log(redis.LOG_WARNING, "filltime " .. fill_time)
--redis.log(redis.LOG_WARNING, "ttl " .. ttl)

## 获取剩余的令牌数
local last_tokens = tonumber(redis.call("get", tokens_key))
if last_tokens == nil then
  last_tokens = capacity
end
--redis.log(redis.LOG_WARNING, "last_tokens " .. last_tokens)

## 获取令牌最后填充时间
local last_refreshed = tonumber(redis.call("get", timestamp_key))
if last_refreshed == nil then
  last_refreshed = 0
end
--redis.log(redis.LOG_WARNING, "last_refreshed " .. last_refreshed)

local delta = math.max(0, now-last_refreshed)
## 计算得到剩余的令牌数
local filled_tokens = math.min(capacity, last_tokens+(delta*rate))
## 大于请求消耗令牌 allowed 设为true
local allowed = filled_tokens >= requested
local new_tokens = filled_tokens
local allowed_num = 0
if allowed then
  new_tokens = filled_tokens - requested
  allowed_num = 1
end

--redis.log(redis.LOG_WARNING, "delta " .. delta)
--redis.log(redis.LOG_WARNING, "filled_tokens " .. filled_tokens)
--redis.log(redis.LOG_WARNING, "allowed_num " .. allowed_num)
--redis.log(redis.LOG_WARNING, "new_tokens " .. new_tokens)

redis.call("setex", tokens_key, ttl, new_tokens)
redis.call("setex", timestamp_key, ttl, now)

return { allowed_num, new_tokens }