今天我们来学习下 `Srping Cloud Gateway` 的全局过滤器 `GloableFilter`。 ## 全局过滤器全局过滤器作用于所有的路由，不需要单独配置，我们可以用它来实现很多统一化处理的业务需求，比如权限认证，IP访问限制等等。 GlobalFilter 接口 ```java package org.springframework.cloud.gateway.filter; import org.springframework.web.server.ServerWebExchange; import reactor.core.publisher.Mono; public interface GlobalFilter { Mono filter(ServerWebExchange exchange, GatewayFilterChain chain); } ``` 我们单独定义也挺简单的，只需要实现`GlobalFilter`, `Ordered`这两个接口就可以了。官网上给的案例如下： ```java public class CustomGlobalFilter implements GlobalFilter, Ordered { @Override public Mono filter(ServerWebExchange exchange, GatewayFilterChain chain) { log.info("custom global filter"); return chain.filter(exchange); } @Override public int getOrder() { return -1; } } ``` 下面我们观察一个自定义的全局过滤器案例： ```java /** * 网关鉴权 * * @author ezhang */ @Component public class AuthFilter implements GlobalFilter, Ordered { private static final Logger log = LoggerFactory.getLogger(AuthFilter.class); private final static long EXPIRE_TIME = Constants.TOKEN_EXPIRE * 60; // 排除过滤的 uri 地址，nacos自行添加 @Autowired private IgnoreWhiteProperties ignoreWhite; @Resource(name = "stringRedisTemplate") private ValueOperations sops; @Value("${ADMIN-IP}") private String adminIp; @Autowired private RedisService redisService; @Override public Mono filter(ServerWebExchange exchange, GatewayFilterChain chain) { String url = exchange.getRequest().getURI().getPath(); String userIpaddr = exchange.getRequest().getRemoteAddress().getAddress().getHostAddress(); // 跳过不需要验证的路径 if (StringUtils.matches(url, ignoreWhite.getWhites())) { return chain.filter(exchange); } String token = getToken(exchange.getRequest()); if (StringUtils.isBlank(token)) { return setUnauthorizedResponse(exchange, "令牌不能为空", com.erow.base.common.constant.HttpStatus.ERROR); } String userStr = sops.get(getTokenKey(token)); if (StringUtils.isNull(userStr)) { return setUnauthorizedResponse(exchange, "登录状态已过期",com.erow.base.common.constant.HttpStatus.FORBIDDEN); } JSONObject obj = JSONObject.parseObject(userStr); String userid = obj.getString("userid"); String username = obj.getString("username"); String rootin = obj.getString("rootin"); String passwordUpdateDate = obj.getString("passwordUpdateDate"); String roleIds = obj.getString("roleIds"); if (StringUtils.isBlank(userid) || StringUtils.isBlank(username)) { return setUnauthorizedResponse(exchange, "令牌验证失败",com.erow.base.common.constant.HttpStatus.ERROR); } //校验用户密码是否已过期 SimpleDateFormat sdf = new SimpleDateFormat("yyyy/MM/dd"); Date parse = null; try { parse = sdf.parse(passwordUpdateDate); int differentDays = DateUtils.daysBetween(parse,new Date()); //修改密码接口放行 if(UserConstants.MAXEXPIRED<= differentDays && !url.equals("/auth/updateUser")){ return setUnauthorizedResponse(exchange, "您的密码已过期,请重新修改密码",com.erow.base.common.constant.HttpStatus.PWD_EXPIRED); } } catch (NullPointerException e) { //修改密码接口放行 if(parse==null && !url.equals("/auth/updateUser")){ return setUnauthorizedResponse(exchange, "首次登录,请先修改密码！",com.erow.base.common.constant.HttpStatus.PWD_EXPIRED); } }catch (Exception e) { e.printStackTrace(); } // 设置过期时间 redisService.expire(getTokenKey(token), EXPIRE_TIME); // 设置用户信息到请求 ServerHttpRequest mutableReq = exchange.getRequest().mutate().header(CacheConstants.DETAILS_USER_ID, userid) .header(CacheConstants.DETAILS_USERNAME, ServletUtils.urlEncode(username)).build(); ServerWebExchange mutableExchange = exchange.mutate().request(mutableReq).build(); return chain.filter(mutableExchange); } private Mono setUnauthorizedResponse(ServerWebExchange exchange, String msg,String code) { ServerHttpResponse response = exchange.getResponse(); response.getHeaders().setContentType(MediaType.APPLICATION_JSON); response.setStatusCode(HttpStatus.INTERNAL_SERVER_ERROR); log.error("[鉴权异常处理]请求路径:{}", exchange.getRequest().getPath()); return response.writeWith(Mono.fromSupplier(() -> { DataBufferFactory bufferFactory = response.bufferFactory(); return bufferFactory.wrap(JSON.toJSONBytes(AjaxResult.error(code,msg))); })); } private String getTokenKey(String token) { return CacheConstants.LOGIN_TOKEN_KEY + token; } private String getLoginUserKey(String rootin,String userId) { return rootin+"_"+CacheConstants.LOGIN_USER_KEY + userId; } /** * 获取请求token */ private String getToken(ServerHttpRequest request) { String token = request.getHeaders().getFirst(CacheConstants.HEADER); if (StringUtils.isNotEmpty(token) && token.startsWith(CacheConstants.TOKEN_PREFIX)) { token = token.replace(CacheConstants.TOKEN_PREFIX, ""); } return token; } @Override public int getOrder() { return -200; } } ``` 上面的自定义全局过滤器中，我们用来统一处理了一些业务需求，比如 token 令牌验证、白名单跳过等等。 ```java @Override public int getOrder() { return -200; } ``` 这部分代码定义过滤器执行的优先级 >order 越大，优先级越低 ## 局部过滤器全局过滤器作用于所有的路由，不需要单独配置。局部过滤器就不一样了，需要单独配置。局部过滤器步骤： 1. 需要继承 `AbstractGatewayFilterFactory`，覆盖相关的方法。 2. 在配置文件中进行配置，如果不配置则不启用此过滤器规则。 ```java @Component public class BlackListUrlFilter extends AbstractGatewayFilterFactory { @Override public GatewayFilter apply(Config config) { return (exchange, chain) -> { String url = exchange.getRequest().getURI().getPath(); if (config.matchBlacklist(url)) { ServerHttpResponse response = exchange.getResponse(); response.getHeaders().add("Content-Type", "application/json;charset=UTF-8"); return exchange.getResponse().writeWith( Mono.just(response.bufferFactory().wrap(JSON.toJSONBytes("请求地址不允许访问")))); } return chain.filter(exchange); }; } public BlackListUrlFilter() { super(Config.class); } public static class Config { private List blacklistUrl; private List blacklistUrlPattern = new ArrayList<>(); public boolean matchBlacklist(String url) { return blacklistUrlPattern.isEmpty() ? false : blacklistUrlPattern.stream().filter(p -> p.matcher(url).find()).findAny().isPresent(); } public List getBlacklistUrl() { return blacklistUrl; } public void setBlacklistUrl(List blacklistUrl) { this.blacklistUrl = blacklistUrl; this.blacklistUrlPattern.clear(); this.blacklistUrl.forEach(url -> { this.blacklistUrlPattern.add(Pattern.compile(url.replaceAll("\*\*", "(.*?)"), Pattern.CASE_INSENSITIVE)); }); } } } ``` 配置文件中配置： ```yml spring: cloud: gateway: routes: # 系统模块 - id: cloud-system uri: lb://cloud-system predicates: - Path=/system/** filters: - StripPrefix=1 - name: BlackListUrlFilter args: blacklistUrl: - /user/list ``` 这样，我们就定义好了一个局部黑名单过滤器。访问 `/user/list` 地址时会返回：请求地址不允许访问。 ## 白名单配置其实在上面的 `AuthFilter` 中已经处理了白名单地址，就是不需要进行校验的地址。 ```java // 跳过不需要验证的路径 if (StringUtils.matches(url, ignoreWhite.getWhites())) { return chain.filter(exchange); } ``` `ignoreWhite` 就是直接读取 yml 中的自定义配置。 ```java @Configuration @RefreshScope @ConfigurationProperties(prefix = "ignore") public class IgnoreWhiteProperties { /** * 放行白名单配置，网关不校验此处的白名单 */ private List whites = new ArrayList<>(); public List getWhites() { return whites; } public void setWhites(List whites) { this.whites = whites; } } ``` yml 配置如下 ```yml # 不校验白名单 ignore: whites: - /auth/logout - /auth/login - /*/v2/api-docs ``` 包含这些地址的 URL 就不需要校验 token 了，直接放行。