[feature] Best practices for JVM threshold monitoring

发现一个jvm监控的现象，我们提供了很多监控指标，但从文档上略显偏薄。是否可以考虑文档中补充：1.实际生产中需要监控哪些指标，阀值设置为多少（or 范围）合适，比如我们一般想知道啥时候增加xmx。2.当指标异常时，是否有对应的配置项进行异常指标的降低，即指标与配置的对应关系。

Start by locating the existing JVM monitoring documentation and the definitions of its metrics and related configuration options. Document which production JVM indicators to watch, suggested thresholds or ranges, when an Xmx increase may be appropriate, and how abnormal indicators map to settings. Done means the guidance and metric-to-configuration relationships are added clearly to the documentation.

Written by the indexing model from the issue text.