#2997 osaf: etcd3.plugin should detect if etcd is down

[Gary Lee]

Currently, write access to the consensus service is checked only when a node enters or leaves the cluster. We should enhance the plugin so it checks for write access to the service regularly, in the event that the etcd server becomes inaccessible without cluster membership changes. If write access is lost, the active SC should consider self-fencing itself.

[Gary Lee]

will be done as part of [#2996]