Adjust keepalive health printing and ordering

node-daemon/pvcnoded/objects/MonitoringInstance.py

@@ -365,7 +365,10 @@ class MonitoringInstance(object):
                 plugin_results.append(future.result())
 
         for result in sorted(plugin_results, key=lambda x: x.plugin_name):
-            if self.config["log_keepalive_plugin_details"]:
+            if (
+                self.config["log_keepalives"]
+                and self.config["log_keepalive_plugin_details"]
+            ):
                 self.logger.out(
                     result.message + f" [-{result.health_delta}]",
                     state="t",
@@ -376,13 +379,6 @@ class MonitoringInstance(object):
         if total_health < 0:
             total_health = 0
 
-        if total_health > 90:
-            health_colour = self.logger.fmt_green
-        elif total_health > 50:
-            health_colour = self.logger.fmt_yellow
-        else:
-            health_colour = self.logger.fmt_red
-
         self.zkhandler.write(
             [
                 (
@@ -391,10 +387,6 @@ class MonitoringInstance(object):
                 ),
             ]
         )
-        self.logger.out(
-            f"Node health: {health_colour}{total_health}%{self.logger.fmt_end}",
-            state="t",
-        )
 
     def run_cleanup(self, plugin):
         return plugin.cleanup()

node-daemon/pvcnoded/objects/NodeInstance.py

@@ -67,6 +67,7 @@ class NodeInstance(object):
         self.network_list = []
         self.domain_list = []
         # Node resources
+        self.health = 100
         self.domains_count = 0
         self.memused = 0
         self.memfree = 0
@@ -224,6 +225,28 @@ class NodeInstance(object):
                         )
                         self.flush_thread.start()
 
+        @self.zkhandler.zk_conn.DataWatch(
+            self.zkhandler.schema.path("node.monitoring.health", self.name)
+        )
+        def watch_node_health(data, stat, event=""):
+            if event and event.type == "DELETED":
+                # The key has been deleted after existing before; terminate this watcher
+                # because this class instance is about to be reaped in Daemon.py
+                return False
+
+            try:
+                data = data.decode("ascii")
+            except AttributeError:
+                data = 100
+
+            try:
+                data = int(data)
+            except ValueError:
+                pass
+
+            if data != self.health:
+                self.health = data
+
         @self.zkhandler.zk_conn.DataWatch(
             self.zkhandler.schema.path("node.memory.free", self.name)
         )