Merge remote-tracking branch 'upstream/master' into kubelet-eviction

Author: TheRealNoob

Makefile

@@ -90,7 +90,7 @@ clean:
 
 .PHONY: test
 test: $(PROMTOOL_BIN) prometheus_alerts.yaml prometheus_rules.yaml
-	@$(PROMTOOL_BIN) test rules tests.yaml
+	@$(PROMTOOL_BIN) test rules tests/*.yaml
 
 $(BIN_DIR):
 	mkdir -p $(BIN_DIR)

alerts/kubelet.libsonnet

@@ -81,14 +81,16 @@ local utils = import '../lib/utils.libsonnet';
             // Some node has a capacity of 1 like AWS's Fargate and only exists while a pod is running on it.
             // We have to ignore this special node in the KubeletTooManyPods alert.
             expr: |||
-              count by (%(clusterLabel)s, node) (
-                (kube_pod_status_phase{%(kubeStateMetricsSelector)s, phase="Running"} == 1)
-                * on (%(clusterLabel)s, namespace, pod) group_left (node)
-                group by (%(clusterLabel)s, namespace, pod, node) (
-                  kube_pod_info{%(kubeStateMetricsSelector)s}
+              (
+                max by (%(clusterLabel)s, instance) (
+                  kubelet_running_pods{%(kubeletSelector)s} > 1
+                )
+                * on (%(clusterLabel)s, instance) group_left(node)
+                max by (%(clusterLabel)s, instance, node) (
+                  kubelet_node_name{%(kubeletSelector)s}
                 )
               )
-              /
+              / on (%(clusterLabel)s, node) group_left()
               max by (%(clusterLabel)s, node) (
                 kube_node_status_capacity{%(kubeStateMetricsSelector)s, resource="pods"} != 1
               ) > 0.95

dashboards/dashboards.libsonnet

@@ -6,4 +6,5 @@
 (import 'scheduler.libsonnet') +
 (import 'proxy.libsonnet') +
 (import 'kubelet.libsonnet') +
-(import 'defaults.libsonnet')
+(import 'defaults.libsonnet') +
+(import 'windows.libsonnet')

dashboards/windows.libsonnet

@@ -72,8 +72,7 @@ local var = g.dashboard.variable;
       )
       + var.query.generalOptions.withLabel('instance')
       + var.query.refresh.onTime()
-      + var.query.generalOptions.showOnDashboard.withLabelAndValue()
-      + var.query.selectionOptions.withMulti(true),
+      + var.query.generalOptions.showOnDashboard.withLabelAndValue(),
 
     namespace:
       var.query.new('namespace')
@@ -124,7 +123,7 @@ local var = g.dashboard.variable;
         statPanel(
           'CPU Utilisation',
           'none',
-          '1 - avg(rate(windows_cpu_time_total{%(clusterLabel)s="$cluster", %(windowsExporterSelector)s, mode="idle"}[1m]))' % $._config
+          '1 - avg(rate(windows_cpu_time_total{%(clusterLabel)s="$cluster", %(windowsExporterSelector)s, mode="idle"}[%(grafanaIntervalVar)s]))' % $._config
         )
         + stat.gridPos.withW(4)
         + stat.gridPos.withH(3),
@@ -830,13 +829,13 @@ local var = g.dashboard.variable;
         + tsPanel.queryOptions.withTargets([
           prometheus.new(
             '${datasource}',
-            'sort_desc(sum by (container) (rate(windows_container_network_received_bytes_total{%(clusterLabel)s="$cluster", namespace="$namespace", pod="$pod"}[1m])))' % $._config
+            'sort_desc(sum by (container) (rate(windows_container_network_received_bytes_total{%(clusterLabel)s="$cluster", namespace="$namespace", pod="$pod"}[%(grafanaIntervalVar)s])))' % $._config
           )
           + prometheus.withLegendFormat('Received : {{ container }}'),
 
           prometheus.new(
             '${datasource}',
-            'sort_desc(sum by (container) (rate(windows_container_network_transmitted_bytes_total{%(clusterLabel)s="$cluster", namespace="$namespace", pod="$pod"}[1m])))' % $._config
+            'sort_desc(sum by (container) (rate(windows_container_network_transmitted_bytes_total{%(clusterLabel)s="$cluster", namespace="$namespace", pod="$pod"}[%(grafanaIntervalVar)s])))' % $._config
           )
           + prometheus.withLegendFormat('Transmitted : {{ container }}'),
         ]),
@@ -1025,19 +1024,19 @@ local var = g.dashboard.variable;
         + tsPanel.queryOptions.withTargets([
           prometheus.new(
             '${datasource}',
-            'max(rate(windows_logical_disk_read_bytes_total{%(clusterLabel)s="$cluster", %(windowsExporterSelector)s, instance="$instance"}[2m]))' % $._config
+            'max(rate(windows_logical_disk_read_bytes_total{%(clusterLabel)s="$cluster", %(windowsExporterSelector)s, instance="$instance"}[%(grafanaIntervalVar)s]))' % $._config
           )
           + prometheus.withLegendFormat('read'),
 
           prometheus.new(
             '${datasource}',
-            'max(rate(windows_logical_disk_write_bytes_total{%(clusterLabel)s="$cluster", %(windowsExporterSelector)s, instance="$instance"}[2m]))' % $._config
+            'max(rate(windows_logical_disk_write_bytes_total{%(clusterLabel)s="$cluster", %(windowsExporterSelector)s, instance="$instance"}[%(grafanaIntervalVar)s]))' % $._config
           )
           + prometheus.withLegendFormat('written'),
 
           prometheus.new(
             '${datasource}',
-            'max(rate(windows_logical_disk_read_seconds_total{%(clusterLabel)s="$cluster", %(windowsExporterSelector)s,  instance="$instance"}[2m]) + rate(windows_logical_disk_write_seconds_total{%(clusterLabel)s="$cluster", %(windowsExporterSelector)s, instance="$instance"}[2m]))' % $._config
+            'max(rate(windows_logical_disk_read_seconds_total{%(clusterLabel)s="$cluster", %(windowsExporterSelector)s,  instance="$instance"}[%(grafanaIntervalVar)s]) + rate(windows_logical_disk_write_seconds_total{%(clusterLabel)s="$cluster", %(windowsExporterSelector)s, instance="$instance"}[%(grafanaIntervalVar)s]))' % $._config
           )
           + prometheus.withLegendFormat('io time'),
         ])

rules/rules.libsonnet

@@ -2,4 +2,5 @@
 (import 'apps.libsonnet') +
 (import 'kube_scheduler.libsonnet') +
 (import 'node.libsonnet') +
-(import 'kubelet.libsonnet')
+(import 'kubelet.libsonnet') +
+(import 'windows.libsonnet')