Reduce the thread overhead in NodeMonitorUpdater.

Before, if N nodes connected quickly, this would spawn N threads all
sleeping for 5 seconds before realizing they don't need to do anything.

core/src/main/java/hudson/node_monitors/NodeMonitorUpdater.java

@@ -4,9 +4,11 @@ import hudson.Extension;
 import hudson.model.Computer;
 import hudson.model.TaskListener;
 import hudson.slaves.ComputerListener;
+import hudson.util.Futures;
 import jenkins.model.Jenkins;
 
 import java.io.IOException;
+import java.util.concurrent.Future;
 import java.util.concurrent.TimeUnit;
 import java.util.concurrent.atomic.AtomicInteger;
 
@@ -20,7 +22,16 @@ import jenkins.util.Timer;
 @Extension
 public class NodeMonitorUpdater extends ComputerListener {
 
-    private final AtomicInteger id = new AtomicInteger();
+    private static final Runnable MONITOR_UPDATER = new Runnable() {
+        @Override
+        public void run() {
+            for (NodeMonitor nm : Jenkins.getInstance().getComputer().getMonitors()) {
+                nm.triggerUpdate();
+            }
+        }
+    };
+
+    private Future<?> future = Futures.precomputed(null);
 
     /**
      * Triggers the update with 5 seconds quiet period, to avoid triggering data check too often
@@ -28,18 +39,9 @@ public class NodeMonitorUpdater extends ComputerListener {
      */
     @Override
     public void onOnline(Computer c, TaskListener listener) throws IOException, InterruptedException {
-        Timer.get().schedule(new Runnable() {
-            final int _id = id.incrementAndGet();
-
-            public void run() {
-                if (id.get() != _id)
-                    return;
-
-                for (NodeMonitor nm : Jenkins.getInstance().getComputer().getMonitors()) {
-                    nm.triggerUpdate();
-                }
-            }
-        }, 5, TimeUnit.SECONDS);
+        synchronized(this) {
+            future.cancel(false);
+            future = Timer.get().schedule(MONITOR_UPDATER, 5, TimeUnit.SECONDS);
+        }
     }
-
 }