参考Spring AOP理解与研发使用_aop开发-CSDN博客进行学习总结
场景分析
假设为了更好的分析某服务的一些接口处理和迭代变动,我们需要对其内部很多thrift接口进行干预处理。我们可以牵起将相关的thrift接口统计出来,然后将其中的各接口进行一套模版的配置用于控制分析处理,在配置上针对各个接口可以定义其基本的一些策略:
透传策略-该接口按实际调用方式触发实际内容;
异常策略-该接口调用时直接抛出异常即可;
mock策略-将具体结果按我们定义的内容进行返回处理;
在实际处理上收敛各个接口从而分析变动点以及不影响实际其他业务等,具体流程图如下。
该方式内部也可以使用切面处理,处于考虑,我们也可以通过过滤器方式来完成,本次按该方式代码如下(个人觉得该场景就应该使用过滤器实现即可),后期按切面处理。
相关代码展示
package org.zyf.javabasic.aop.mock;
import com.alibaba.fastjson.JSON;
import com.dianping.lion.Environment;
import com.google.common.cache.CacheBuilder;
import com.google.common.cache.CacheLoader;
import com.google.common.cache.LoadingCache;
import com.google.common.collect.Lists;
import com.zyf.horse.common.RpcRole;
import com.zyf.horse.rpc.handler.filter.Filter;
import com.zyf.horse.rpc.handler.filter.FilterHandler;
import com.zyf.horse.rpc.meta.RpcInvocation;
import com.zyf.horse.rpc.meta.RpcResult;
import com.zyf.service.mobile.mtthrift.client.pool.MTThriftPoolConfig;
import com.zyf.service.mobile.mtthrift.proxy.ThriftClientProxy;
import com.zyf.product.bsap.thrift.domain.common.ProductBsapThirftException;
import com.zyf.product.bsap.thrift.domain.mock.MockRuleDto;
import com.zyf.product.bsap.thrift.iface.MockRuleThriftService;
import com.zyf.tsp.product.bsap.enums.MockRuleType;
import com.zyf.tsp.product.bsap.mcc.BsapCommonSwitch;
import lombok.extern.log4j.Log4j2;
import org.apache.commons.collections.CollectionUtils;
import org.apache.commons.lang3.StringUtils;
import org.apache.thrift.TException;
import org.springframework.stereotype.Component;
import java.util.HashSet;
import java.util.List;
import java.util.Set;
import java.util.concurrent.TimeUnit;
import java.util.regex.Pattern;
/**
* @author yanfengzhang
* @description mock组件对指定接口按指定内容进行干预
* @date 2021/12/28 23:33
*/
@Log4j2
@Component
public class MockRemoteThriftClientFilter implements Filter {
private static MockRuleThriftService.Iface mockRuleThriftService = null;
private static final String MOCK_CLIENT_APPKEY = BsapCommonSwitch.getMockClientApp();
private static final String MOCK_SERVER_APPKEY = "com.zyf.product.bsap";
private static volatile Set<String> rpcInfos = new HashSet<>();
static {
MockRemoteThriftClientFilter.init(MOCK_CLIENT_APPKEY);
}
private static LoadingCache<String, List<MockRuleDto>> mockRuleCache = CacheBuilder.newBuilder()
/*初始化加载一次,后续每隔十分钟刷新缓存数据*/
.refreshAfterWrite(BsapCommonSwitch.getMockRuleReflushTime(), TimeUnit.MINUTES)
/*构建缓存*/
.build(new CacheLoader<String, List<MockRuleDto>>() {
/*初始化加载数据的缓存信息*/
@Override
public List<MockRuleDto> load(String str) throws TException, ProductBsapThirftException {
Long total = 60L;
List<MockRuleDto> mockRuleDtos = Lists.newArrayList();
Long validMockRuleCount = mockRuleThriftService.getValidCount();
if (validMockRuleCount < total) {
/*总接口数较少时直接全量拉取*/
return mockRuleThriftService.getAllMockRules();
}
/*数据量大时进行分次拉取*/
for (long i = 0; i <= validMockRuleCount; ) {
mockRuleDtos.addAll(mockRuleThriftService.getMockRulesByRange(i, i + total));
i = i + total + 1;
}
return mockRuleDtos;
}
});
/**
* 初始化thrift客户端
*/
public static void init(String appKey) {
MTThriftPoolConfig mtThriftPoolConfig = new MTThriftPoolConfig();
mtThriftPoolConfig.setMaxActive(5);
mtThriftPoolConfig.setMaxIdle(2);
mtThriftPoolConfig.setMinIdle(1);
mtThriftPoolConfig.setMaxWait(3000L);
mtThriftPoolConfig.setTestOnBorrow(false);
ThriftClientProxy mockClientProxy = new ThriftClientProxy();
mockClientProxy.setMtThriftPoolConfig(mtThriftPoolConfig);
mockClientProxy.setAppKey(appKey);
mockClientProxy.setRemoteAppkey(MOCK_SERVER_APPKEY);
mockClientProxy.setTimeout(5000);
mockClientProxy.setServiceInterface(MockRuleThriftService.class);
mockClientProxy.setFilterByServiceName(true);
try {
mockClientProxy.afterPropertiesSet();
mockRuleThriftService = (MockRuleThriftService.Iface) mockClientProxy.getObject();
} catch (Exception e) {
throw new RuntimeException(e);
}
}
private static List<MockRuleDto> loadData(String str) throws TException, ProductBsapThirftException {
Long total = 60L;
List<MockRuleDto> mockRuleDtos = Lists.newArrayList();
Long validMockRuleCount = mockRuleThriftService.getValidCount();
if (validMockRuleCount < total) {
/*总接口数较少时直接全量拉取*/
return mockRuleThriftService.getAllMockRules();
}
/*数据量大时进行分次拉取*/
for (long i = 0; i <= validMockRuleCount; ) {
mockRuleDtos.addAll(mockRuleThriftService.getMockRulesByRange(i, i + total));
i = i + total + 1;
}
return mockRuleDtos;
}
/**
* 将此过滤进行优先处理,优先级至高
*
* @return 过滤器优先级优先级
*/
@Override
public int getPriority() {
return 1000;
}
/**
* 本sdk作为本身的调用服务进行处理
*
* @return
*/
@Override
public RpcRole getRole() {
return RpcRole.INVOKER;
}
/**
* 实时拦截干预处理
*
* @param rpcInvocation rpc调用消息
* @param filterHandler 过滤处理
* @return 过滤处理结果
* @throws Throwable 异常消息
*/
@Override
public RpcResult filter(RpcInvocation rpcInvocation, FilterHandler filterHandler) throws Throwable {
final String MOCK_RULE_DATA = "mock_data";
/*1.特定service不进行干预*/
if (BsapCommonSwitch.getCommonRpcAppKeyList().contains(rpcInvocation.getServiceInterface().getName())) {
return filterHandler.handle(rpcInvocation);
}
/*特定服务进行记录相关内容*/
if (BsapCommonSwitch.getMockRpcInfo() && Environment.getAppName().equals(BsapCommonSwitch.getMockClientApp())) {
rpcInfos.add(rpcInvocation.getServiceInterface().getName() + ":" + rpcInvocation.getMethod().getName());
FileWriteUtils.writeToFile(rpcInfos.toString(), "/var/zyf/rpcInfo.txt");
}
/*2.无mock规则直接按原方式进行处理*/
List<MockRuleDto> mockRuleDtos = mockRuleCache.getUnchecked(MOCK_RULE_DATA);
if (CollectionUtils.isEmpty(mockRuleDtos)) {
/*全新修改0516:如果不在mock配置中都返回默认数据干预*/
return defaultDealResult(rpcInvocation);
}
/*3.查看当前客户端调用接口是否在mock规则的配置中*/
MockRuleDto mockRuleDto = mockRuleDtos.stream()
.filter(mockRuleDtoInfo -> mockRuleDtoInfo.getInterfaceInfo().equals(rpcInvocation.getServiceInterface().getName())
&& mockRuleDtoInfo.getMethodName().equals(rpcInvocation.getMethod().getName()))
.findAny().orElse(null);
/*4.如果对应接口和方法并不是指定的mock数据,直接按原方式进行处理(后续改成不在配置中的直接抛错)*/
if (null == mockRuleDto) {
/*全新修改0516:如果不在mock配置中都返回默认数据干预*/
return defaultDealResult(rpcInvocation);
}
/*5.查看当前的mock规则,如果是透传则直接进行处理*/
int isMock = mockRuleDto.getIsMock();
if (MockRuleType.PENETRATE.getType() == isMock) {
/*如果是透传,则直接按原方式进行处理*/
return filterHandler.handle(rpcInvocation);
}
/*6.如果是拒绝类型则直接抛出异常*/
if (MockRuleType.EWJECT.getType() == isMock) {
/*如果是拒绝,则直接按异常返回*/
throw new Exception("当前rpc接口mock规则为拒绝策略,不允许远程调用!");
}
/*7.需要进行mock的,按当前的处理方式进行mock数据的返回*/
/*返回的具体类*/
String resultTypeName = rpcInvocation.getMethod().getAnnotatedReturnType().getType().getTypeName();
if (resultTypeName.equals("void") && mockRuleDto.getResult().equals("void")) {
return new RpcResult(true);
}
if (mockRuleDto.getResult().equals("true") && (resultTypeName.equals("boolean") || resultTypeName.equals("java.lang.Boolean"))) {
return new RpcResult(true);
}
if (isNumeric(mockRuleDto.getResult()) && (resultTypeName.equals("int") || resultTypeName.equals("java.lang.Integer"))) {
return new RpcResult(Integer.valueOf(mockRuleDto.getResult()));
}
if (isNumeric(mockRuleDto.getResult()) && (resultTypeName.equals("long") || resultTypeName.equals("java.lang.Long"))) {
return new RpcResult(Long.valueOf(mockRuleDto.getResult()));
}
if (isDoubleOrFloat(mockRuleDto.getResult()) && (resultTypeName.equals("double") || resultTypeName.equals("java.lang.Double"))) {
return new RpcResult(Double.valueOf(mockRuleDto.getResult()));
}
if (isDoubleOrFloat(mockRuleDto.getResult()) && (resultTypeName.equals("float") || resultTypeName.equals("java.lang.Float"))) {
return new RpcResult(Float.valueOf(mockRuleDto.getResult()));
}
Object rpcResult = JSON.parseObject(mockRuleDto.getResult(), rpcInvocation.getMethod().getReturnType());
return new RpcResult(rpcResult);
}
private boolean isNumeric(String str) {
if (StringUtils.isBlank(str)) {
return false;
}
for (int i = str.length(); --i >= 0; ) {
int chr = str.charAt(i);
if (chr < 48 || chr > 57) {
return false;
}
}
return true;
}
/**
* 是否为浮点数?double或float类型
*
* @param str 传入的字符串
* @return 是浮点数返回true, 否则返回false
*/
public static boolean isDoubleOrFloat(String str) {
Pattern pattern = Pattern.compile("^[-\\+]?[.\\d]*$");
return pattern.matcher(str).matches();
}
/**
* 没有mock的直接按默认的来处理
*
* @param rpcInvocation
* @return
*/
private RpcResult defaultDealResult(RpcInvocation rpcInvocation) {
String resultTypeName = rpcInvocation.getMethod().getAnnotatedReturnType().getType().getTypeName();
if (resultTypeName.equals("void")) {
return new RpcResult(true);
}
if (resultTypeName.equals("boolean") || resultTypeName.equals("java.lang.Boolean")) {
return new RpcResult(true);
}
if (resultTypeName.equals("int") || resultTypeName.equals("java.lang.Integer")) {
return new RpcResult(0);
}
if (resultTypeName.equals("long") || resultTypeName.equals("java.lang.Long")) {
return new RpcResult(0L);
}
if (resultTypeName.equals("double") || resultTypeName.equals("java.lang.Double")) {
return new RpcResult(0D);
}
if (resultTypeName.equals("float") || resultTypeName.equals("java.lang.Float")) {
return new RpcResult(0F);
}
return null;
}
}
————————————————
版权声明:本文为博主原创文章,遵循 CC 4.0 BY-SA 版权协议,转载请附上原文出处链接和本声明。
原文链接:https://blog.csdn.net/xiaofeng10330111/article/details/106086455