关于Feign降级导致seata服务不能回滚问题

- [ ] I have searched the [issues](https://github.com/seata/seata/issues) of this repository and believe that this is not a duplicate.

### Ⅰ. Issue Description
#2088
参考了上面这个问题的相关做法，但还是有问题：
假如现在有a、b、c三个服务，在a服务里顺序调用b、c服务，b、c服务都配置了降级方法，并且在降级方法里配置了回滚：GlobalTransactionContext.reload(RootContext.getXID()).rollback();
@wangwei-ying
现在是在调用b服务时超时了，可能是因为异常信息被feign处理掉了，seata感知不到，这时会继续调用c服务，但是在c服务里面尝试获取全局事务ID：RootContext.getXID(); 是获取不到的，感觉是没注册到，所以就导致c服务里面的事务自己本地提交了，而a、b服务的事务都回滚了

### Ⅱ. Describe what happened

```
```

### Ⅲ. Describe what you expected to happen
有什么办法能让c服务不被继续调用到，或者能让c服务的事务被回滚呢？

### Ⅳ. How to reproduce it (as minimally and precisely as possible)

### Ⅴ. Anything else we need to know?

### Ⅵ. Environment:

- JDK version :1.8
- OS :windoes 10
- Others: seata1.1.0

Start with the Feign fallback flow described for services a, b, and c, and inspect how RootContext.getXID() and GlobalTransactionContext.reload(...).rollback() behave in Seata 1.1.0. Reproduce the timeout when b is called from a, then verify whether c is invoked without a global transaction and whether its local transaction commits. Done means the failure and a reliable way to prevent c or roll it back are documented.

Written by the indexing model from the issue text.