Heml deployment adds probe detection to prevent cluster splitting caused by rolling updates

[kuaile-zc]

Problem description:
When we used Heml deplyment nacos in K8s, we found that rolling updates can cause a number of problems such as cluster splitting.
So we analyze the reasons, we find the following rules.
Because the rolling update does not add probe verification, a short period of time during the update will cause all the PODS to be suspended !
We added probes and tested the method successfully in a K8S environment ,So we would like to ask if we can come up with PR to help all community partners avoid this problem !

See:
1.Snapshots taken when the cluster status is abnormal may cause Service MetaData and Instance data synchronization problems.

2. PR Add startupProbe and livenessProbe #269
   3.We plan to communicate if nacos add cluster health check access
   
   In order to increase self-healing ability

当我们在k8中使用Heml部署nacos时，我们发现滚动更新会导致许多问题，比如集群分裂。
所以我们分析原因，我们发现以下规则。
由于滚动更新没有添加探针验证，更新过程中很短的一段时间将导致所有pod杀死!
我们在K8S环境下添加了探头并成功测试了该方法，所以我们想问一下，我们是否可以想出PR来帮助所有社区合作伙伴避免这个问题!

如下:

1. 在集群状态异常时进行快照，可能会导致“服务元数据”和“实例数据同步”出现问题。
2. PR Add startupProbe and livenessProbe #269
3. 如果naos添加集群健康检查访问，我们计划进行沟通

[kuaile-zc]

related issues：
#214

[kuaile-zc]

To see if we can create a new interface to check if the current Pod is in a normal cluster state, we used livenessProbe to check this interface, which solved JRaft's inability to heal when all instances were killed.

[kuaile-zc]

@paderlol