Merge pull request #161 from splunk/nodesAndPodsUpstream

Some tweaks to Nodes and Pods dashboard

Author: metalmatze

dashboards/node.libsonnet

@@ -20,7 +20,8 @@ local gauge = promgrafonnet.gauge;
         )
         .addTarget(prometheus.target('max(node_load1{%(clusterLabel)s="$cluster", %(nodeExporterSelector)s, instance="$instance"})' % $._config, legendFormat='load 1m'))
         .addTarget(prometheus.target('max(node_load5{%(clusterLabel)s="$cluster", %(nodeExporterSelector)s, instance="$instance"})' % $._config, legendFormat='load 5m'))
-        .addTarget(prometheus.target('max(node_load15{%(clusterLabel)s="$cluster", %(nodeExporterSelector)s, instance="$instance"})' % $._config, legendFormat='load 15m'));
+        .addTarget(prometheus.target('max(node_load15{%(clusterLabel)s="$cluster", %(nodeExporterSelector)s, instance="$instance"})' % $._config, legendFormat='load 15m'))
+        .addTarget(prometheus.target('count(node_cpu_seconds_total{%(clusterLabel)s="$cluster", %(nodeExporterSelector)s, instance="$instance", mode="user"})' % $._config, legendFormat='logical cores'));
 
       local cpuByCore =
         graphPanel.new(

dashboards/pods.libsonnet

@@ -37,6 +37,10 @@ local numbersinglestat = promgrafonnet.numbersinglestat;
           'sum by(container) (kube_pod_container_resource_limits{%(kubeStateMetricsSelector)s, %(clusterLabel)s="$cluster", namespace="$namespace", resource="memory", pod="$pod", container=~"$container"})' % $._config,
           legendFormat='Limit: {{ container }}',
         ))
+        .addTarget(prometheus.target(
+          'sum by(container_name) (container_memory_cache{%(cadvisorSelector)s, namespace="$namespace", pod_name=~"$pod", container_name=~"$container", container_name!="POD"})' % $._config,
+          legendFormat='Cache: {{ container_name }}',
+        ))
       );
 
       local cpuRow = row.new()
@@ -80,7 +84,30 @@ local numbersinglestat = promgrafonnet.numbersinglestat;
         )
         .addTarget(prometheus.target(
           'sort_desc(sum by (pod_name) (rate(container_network_receive_bytes_total{%(cadvisorSelector)s, %(clusterLabel)s="$cluster", namespace="$namespace", pod_name="$pod"}[1m])))' % $._config,
-          legendFormat='{{ pod_name }}',
+          legendFormat='RX: {{ pod_name }}',
+        ))
+        .addTarget(prometheus.target(
+          'sort_desc(sum by (pod_name) (rate(container_network_transmit_bytes_total{%(cadvisorSelector)s, %(clusterLabel)s="$cluster", namespace="$namespace", pod_name="$pod"}[1m])))' % $._config,
+          legendFormat='TX: {{ pod_name }}',
+        ))
+      );
+
+      local restartsRow = row.new()
+                          .addPanel(
+        graphPanel.new(
+          'Total Restarts Per Container',
+          datasource='$datasource',
+          format='short',
+          min=0,
+          span=12,
+          legend_rightSide=true,
+          legend_alignAsTable=true,
+          legend_current=true,
+          legend_avg=true,
+        )
+        .addTarget(prometheus.target(
+          'max by (container) (kube_pod_container_status_restarts_total{%(kubeStateMetricsSelector)s, %(clusterLabel)s="$cluster", namespace="$namespace", pod="$pod", container=~"$container"})' % $._config,
+          legendFormat='Restarts: {{ container }}',
         ))
       );
 
@@ -158,6 +185,7 @@ local numbersinglestat = promgrafonnet.numbersinglestat;
       .addAnnotation(restartAnnotation)
       .addRow(memoryRow)
       .addRow(cpuRow)
-      .addRow(networkRow),
+      .addRow(networkRow)
+      .addRow(restartsRow),
   },
 }