windows: Use instance label for hostname in hostname panel (#1353)

* windows: Use instance label for hostname in hostname panel

* Fix NTP delay units (thanks to linter)

* Add units to panels

* Add none units to cpu count to pass linter

* Update win mixin

common-lib/common/panels/cpu/stat/count.libsonnet

@@ -16,6 +16,7 @@ base {
 
   stylize(allLayers=true):
     (if allLayers then super.stylize() else {})
-    + generic.info.stylize(allLayers=false),
+    + generic.info.stylize(allLayers=false)
+    + g.panel.stat.standardOptions.withUnit('none'),
 
 }

windows-active-directory-mixin/mixin.libsonnet

@@ -3,16 +3,13 @@ local alerts = import './alerts/alerts.libsonnet';
 local g = import './g.libsonnet';
 local var = g.dashboard.variable;
 local activedirectorymixin =
-  windowsobservlib.new(
-    filteringSelector='job=~"integrations/windows_exporter"',
-    uid='active-directory',
-    groupLabels=['job'],
-    instanceLabels=['instance'],
-  )
+  windowsobservlib.new()
 
   {
     config+: {
       enableADDashboard: true,
+      groupLabels: ['job'],
+      uid: 'active-directory',
     },
   }
 

windows-mixin/mixin.libsonnet

@@ -4,11 +4,7 @@ local winlib = import 'windows-observ-lib/main.libsonnet';
 local config = (import 'config.libsonnet')._config;
 {
   local windows =
-    winlib.new(
-      dashboardNamePrefix=config.dashboardNamePrefix,
-      uid=config.uid,
-      filteringSelector=config.filteringSelector,
-    )
+    winlib.new()
     +
     {
       config+: config,

windows-observ-lib/config.libsonnet

@@ -0,0 +1,42 @@
+{
+  // any modular library should include as inputs:
+  // 'dashboardNamePrefix' - Use as prefix for all Dashboards and (optional) rule groups
+  // 'filteringSelector' - Static selector to apply to ALL dashboard variables of type query, panel queries, alerts and recording rules.
+  // 'groupLabels' - one or more labels that can be used to identify 'group' of instances. In simple cases, can be 'job' or 'cluster'.
+  // 'instanceLabels' - one or more labels that can be used to identify single entity of instances. In simple cases, can be 'instance' or 'pod'.
+  // 'uid' - UID to prefix all dashboards original uids
+  groupLabels: ['job'],
+  instanceLabels: ['instance'],
+  filteringSelector: 'job=~".*windows.*"',
+  dashboardTags: ['windows'],
+  uid: 'windows',
+  dashboardNamePrefix: '',
+
+  // optional
+  ignoreVolumes: 'HarddiskVolume.*',
+  alertsCPUThresholdWarning: '90',
+  alertMemoryUsageThresholdCritical: '90',
+  alertDiskUsageThresholdCritical: '90',
+  dashboardPeriod: 'now-1h',
+  dashboardTimezone: 'default',
+  dashboardRefresh: '1m',
+
+  // optional Windows AD
+  alertsHighPendingReplicationOperations: 50,  // count
+  alertsHighReplicationSyncRequestFailures: 0,  // count
+  alertsHighPasswordChanges: 25,  // count
+  alertsMetricsDownJobName: 'integrations/windows_exporter',
+  enableADDashboard: false,
+
+  // logs lib related
+  enableLokiLogs: true,
+  extraLogLabels: ['channel', 'source', 'keywords', 'level'],
+  logsVolumeGroupBy: 'level',
+  showLogsVolume: true,
+  logsExtraFilters:
+    |||
+      | label_format timestamp="{{__timestamp__}}"
+      | drop channel_extracted,source_extracted,computer_extracted,level_extracted,keywords_extracted
+      | line_format `{{ if eq "[[instance]]" ".*" }}{{ alignLeft 25 .instance}}|{{end}}{{alignLeft 12 .channel }}| {{ alignLeft 25 .source}}| {{ .message }}`
+    |||,
+}

windows-observ-lib/main.libsonnet

@@ -1,4 +1,5 @@
 local alerts = import './alerts.libsonnet';
+local config = import './config.libsonnet';
 local dashboards = import './dashboards.libsonnet';
 local datasources = import './datasources.libsonnet';
 local g = import './g.libsonnet';
@@ -7,58 +8,10 @@ local targets = import './targets.libsonnet';
 local commonlib = import 'common-lib/common/main.libsonnet';
 
 {
-  new(
-    filteringSelector,
-    groupLabels=['job'],
-    instanceLabels=['instance'],
-    dashboardNamePrefix='',
-    dashboardTags=[uid],
-    uid,
-  ): {
+  new(): {
 
     local this = self,
-    config: {
-      // any modular library should include as inputs:
-      // 'dashboardNamePrefix' - Use as prefix for all Dashboards and (optional) rule groups
-      // 'filteringSelector' - Static selector to apply to ALL dashboard variables of type query, panel queries, alerts and recording rules.
-      // 'groupLabels' - one or more labels that can be used to identify 'group' of instances. In simple cases, can be 'job' or 'cluster'.
-      // 'instanceLabels' - one or more labels that can be used to identify single entity of instances. In simple cases, can be 'instance' or 'pod'.
-      // 'uid' - UID to prefix all dashboards original uids
-      groupLabels: groupLabels,
-      instanceLabels: instanceLabels,
-      filteringSelector: filteringSelector,
-      dashboardTags: dashboardTags,
-      uid: uid,
-      dashboardNamePrefix: dashboardNamePrefix,
-
-      // optional
-      ignoreVolumes: 'HarddiskVolume.*',
-      alertsCPUThresholdWarning: '90',
-      alertMemoryUsageThresholdCritical: '90',
-      alertDiskUsageThresholdCritical: '90',
-      dashboardPeriod: 'now-1h',
-      dashboardTimezone: 'default',
-      dashboardRefresh: '1m',
-
-      // optional Windows AD
-      alertsHighPendingReplicationOperations: 50,  // count
-      alertsHighReplicationSyncRequestFailures: 0,  // count
-      alertsHighPasswordChanges: 25,  // count
-      alertsMetricsDownJobName: 'integrations/windows_exporter',
-      enableADDashboard: false,
-
-      // logs lib related
-      enableLokiLogs: true,
-      extraLogLabels: ['channel', 'source', 'keywords', 'level'],
-      logsVolumeGroupBy: 'level',
-      showLogsVolume: true,
-      logsExtraFilters:
-        |||
-          | label_format timestamp="{{__timestamp__}}"
-          | drop channel_extracted,source_extracted,computer_extracted,level_extracted,keywords_extracted
-          | line_format `{{ if eq "[[instance]]" ".*" }}{{ alignLeft 25 .instance}}|{{end}}{{alignLeft 12 .channel }}| {{ alignLeft 25 .source}}| {{ .message }}`
-        |||,
-    },
+    config: config,
     grafana: {
       variables: commonlib.variables.new(
         filteringSelector=this.config.filteringSelector,
@@ -74,7 +27,7 @@ local commonlib = import 'common-lib/common/main.libsonnet';
           reboot: commonlib.annotations.reboot.new(
                     title='Reboot',
                     target=this.grafana.targets.reboot,
-                    instanceLabels=std.join(',', instanceLabels),
+                    instanceLabels=std.join(',', this.config.instanceLabels),
                   )
                   + commonlib.annotations.base.withTagKeys(std.join(',', this.config.groupLabels + this.config.instanceLabels)),
         }
@@ -124,5 +77,8 @@ local commonlib = import 'common-lib/common/main.libsonnet';
     },
 
   },
+  withConfigMixin(config): {
+    config+: config,
+  },
 
 }

windows-observ-lib/mixin.libsonnet

@@ -0,0 +1,10 @@
+local windowslib = import './main.libsonnet';
+{
+  _config:: {},
+  _windowsib::
+    windowslib.new()
+    + windowslib.withConfigMixin(self._config),
+  grafanaDashboards+:: self._windowsib.grafana.dashboards,
+  prometheusAlerts+:: self._windowsib.prometheus.alerts,
+  prometheusRules+:: self._windowsib.prometheus.recordingRules,
+}

windows-observ-lib/panels.libsonnet

@@ -187,22 +187,25 @@ local utils = commonlib.utils;
 
             A high number of context switches or interrupts can indicate that the system is overloaded or that there are problems with specific devices or processes.
           |||
-        ),
+        )
+        + g.panel.timeSeries.standardOptions.withUnit('short'),
       systemExceptions:
         commonlib.panels.generic.timeSeries.base.new(
           'System calls and exceptions',
           targets=[
             t.windowsSystemExceptions,
             t.windowsSystemCalls,
           ],
-        ),
+        )
+        + g.panel.timeSeries.standardOptions.withUnit('short'),
       systemThreads:
         commonlib.panels.generic.timeSeries.base.new(
           'System threads',
           targets=[
             t.windowsSystemThreads,
           ],
-        ),
+        )
+        + g.panel.timeSeries.standardOptions.withUnit('short'),
       timeNtpStatus:
         commonlib.panels.system.statusHistory.ntp.new(
           'NTP status',
@@ -224,7 +227,7 @@ local utils = commonlib.utils;
             Time offset: Absolute time offset between the system clock and the chosen time source, in seconds.
           |||
         )
-        + g.panel.timeSeries.standardOptions.withUnit('seconds')
+        + g.panel.timeSeries.standardOptions.withUnit('s')
         + g.panel.timeSeries.standardOptions.withNoValue('No data. Please check that "time" collector is enabled.'),
       cpuCount: commonlib.panels.cpu.stat.count.new(targets=[t.cpuCount]),
       cpuUsageTs: commonlib.panels.cpu.timeSeries.utilization.new(targets=[t.cpuUsage]),
@@ -242,17 +245,19 @@ local utils = commonlib.utils;
           CPU usage by different modes.
         |||
       ),
-      cpuQueue: commonlib.panels.generic.timeSeries.base.new(
-        'CPU average queue size',
-        targets=[t.cpuQueue],
-        description=|||
-          The CPU average queue size in Windows, often referred to as the "Processor Queue Length" or "CPU Queue Length," is a metric that measures the number of threads or tasks waiting to be processed by the central processing unit (CPU) at a given moment.
-          It is an essential performance indicator that reflects the workload and responsiveness of the CPU.
-          When the CPU queue length is high, it indicates that there are more tasks in line for processing than the CPU can handle immediately.
+      cpuQueue:
+        commonlib.panels.generic.timeSeries.base.new(
+          'CPU average queue size',
+          targets=[t.cpuQueue],
+          description=|||
+            The CPU average queue size in Windows, often referred to as the "Processor Queue Length" or "CPU Queue Length," is a metric that measures the number of threads or tasks waiting to be processed by the central processing unit (CPU) at a given moment.
+            It is an essential performance indicator that reflects the workload and responsiveness of the CPU.
+            When the CPU queue length is high, it indicates that there are more tasks in line for processing than the CPU can handle immediately.
 
-          This can lead to system slowdowns, decreased responsiveness, and potential performance issues. High CPU queue lengths are often associated with CPU saturation, where the CPU is struggling to keep up with the demands placed on it.
-        |||
-      ),
+            This can lead to system slowdowns, decreased responsiveness, and potential performance issues. High CPU queue lengths are often associated with CPU saturation, where the CPU is struggling to keep up with the demands placed on it.
+          |||
+        )
+        + g.panel.timeSeries.standardOptions.withUnit('short'),
       memoryTotalBytes: commonlib.panels.memory.stat.total.new(targets=[t.memoryTotalBytes]),
       memoryPageTotalBytes:
         commonlib.panels.memory.stat.total.new(
@@ -363,7 +368,7 @@ local utils = commonlib.utils;
           targets=[t.osInfo],
           description="System's hostname."
         )
-        { options+: { reduceOptions+: { fields: '/^hostname$/' } } },
+        { options+: { reduceOptions+: { fields: '/^instance$/' } } },
       networkErrorsAndDroppedPerSec:
         commonlib.panels.network.timeSeries.errors.new(
           'Network errors and dropped packets',