Hello community: According to ISSUE[1], the more metadata, the slower the matching speed.
The solution of this PR[2] is to add a LRU cache to make the same URL match only once. This LRU cache uses `MemoryLimiter` to solve the memory problem: when the available memory of the JVM is less than a certain value, the cleanup action will be triggered. If you are free, please see the PR[2] and help this PR[2] to solve the problem better. Looking forward to your reply. [1] https://github.com/apache/incubator-shenyu/issues/3177 [2] https://github.com/apache/incubator-shenyu/pull/3245
