If thread launch_cluster_Monitor() and launch_leader_election() operates on the configmap at the same time, Will cause a error 'Exception in thread "Thread-1"'.
This error will cause the thread to get stuck. Configmap will not be updated and the error "data too old" will be reported.
Just passing kubernetes_API exceptions is not enough, all are more appropriate.
Change-Id: I6baa9ece474f9c937fe9bce2231ef500562e0406
6d808ceb47