Hardening of cluster member leaving path, see #3309

* Removed leader commands for Shutdown and Exit
* Member shutdown itself  when it sees itself as Exiting
* Singleton cluster with status Exiting will shutdown itself,
  in case the Exiting gossip never arrives
* Exiting member not part convergence check
* Exiting member is removed by leader (on convergence) when the
  exiting member is in the unreachable set, i.e. sucessfully shutdown
* Reverted the change made for #3266, i.e. Exiting is
  detected as unreachable again.
* Adjust ClusterSingletonManager to new Exiting behaviour
* Fix bug in HeartbeatSender, which caused it to continue to
  send heartbeats to removed nodes, instead of rebalancing
* Refactoring of leaderActions method
* Leaving section in docs

akka-cluster/src/main/scala/akka/cluster/ClusterEvent.scala

@@ -146,8 +146,6 @@ object ClusterEvent {
   case class UnreachableMember(member: Member) extends ClusterDomainEvent
 
   /**
-   * INTERNAL API
-   *
    * Current snapshot of cluster node metrics. Published to subscribers.
    */
   case class ClusterMetricsChanged(nodeMetrics: Set[NodeMetrics]) extends ClusterDomainEvent {
@@ -200,13 +198,6 @@ object ClusterEvent {
 
       val allNewUnreachable = newGossip.overview.unreachable -- oldGossip.overview.unreachable
 
-      val unreachableGroupedByAddress =
-        List(newGossip.overview.unreachable, oldGossip.overview.unreachable).flatten.groupBy(_.uniqueAddress)
-      val unreachableDownMembers = unreachableGroupedByAddress collect {
-        case (_, newMember :: oldMember :: Nil) if newMember.status == Down && newMember.status != oldMember.status ⇒
-          newMember
-      }
-
       val removedMembers = (oldGossip.members -- newGossip.members -- newGossip.overview.unreachable) ++
         (oldGossip.overview.unreachable -- newGossip.overview.unreachable)
       val removedEvents = removedMembers.map(m ⇒ MemberRemoved(m.copy(status = Removed)))