local grafana = import 'github.com/grafana/grafonnet-lib/grafonnet/grafana.libsonnet'; local dashboard = grafana.dashboard; local row = grafana.row; local prometheus = grafana.prometheus; local template = grafana.template; local graphPanel = grafana.graphPanel; local grafana70 = import 'github.com/grafana/grafonnet-lib/grafonnet-7.0/grafana.libsonnet'; local gaugePanel = grafana70.panel.gauge; local table = grafana70.panel.table; { new(config=null, platform=null, uid=null):: { local prometheusDatasourceTemplate = { current: { text: 'default', value: 'default', }, hide: 0, label: 'Data Source', name: 'datasource', options: [], query: 'prometheus', refresh: 1, regex: '', type: 'datasource', }, local clusterTemplatePrototype = template.new( 'cluster', '$datasource', '', hide=if config.showMultiCluster then '' else '2', refresh='time', label='Cluster', ), local clusterTemplate = if platform == 'Darwin' then clusterTemplatePrototype { query: 'label_values(node_uname_info{%(nodeExporterSelector)s, sysname="Darwin"}, %(clusterLabel)s)' % config } else clusterTemplatePrototype { query: 'label_values(node_uname_info{%(nodeExporterSelector)s, sysname!="Darwin"}, %(clusterLabel)s)' % config }, local instanceTemplatePrototype = template.new( 'instance', '$datasource', '', refresh='time', label='Instance', ), local instanceTemplate = if platform == 'Darwin' then instanceTemplatePrototype { query: 'label_values(node_uname_info{%(nodeExporterSelector)s, %(clusterLabel)s="$cluster", sysname="Darwin"}, instance)' % config } else instanceTemplatePrototype { query: 'label_values(node_uname_info{%(nodeExporterSelector)s, %(clusterLabel)s="$cluster", sysname!="Darwin"}, instance)' % config }, local idleCPU = graphPanel.new( 'CPU Usage', datasource='$datasource', span=6, format='percentunit', max=1, min=0, stack=true, ) .addTarget(prometheus.target( ||| ( (1 - sum without (mode) (rate(node_cpu_seconds_total{%(nodeExporterSelector)s, mode=~"idle|iowait|steal", instance="$instance", %(clusterLabel)s="$cluster"}[$__rate_interval]))) / ignoring(cpu) group_left count without (cpu, mode) (node_cpu_seconds_total{%(nodeExporterSelector)s, mode="idle", instance="$instance", %(clusterLabel)s="$cluster"}) ) ||| % config, legendFormat='{{cpu}}', intervalFactor=5, )), local systemLoad = graphPanel.new( 'Load Average', datasource='$datasource', span=6, format='short', min=0, fill=0, ) .addTarget(prometheus.target('node_load1{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}' % config, legendFormat='1m load average')) .addTarget(prometheus.target('node_load5{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}' % config, legendFormat='5m load average')) .addTarget(prometheus.target('node_load15{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}' % config, legendFormat='15m load average')) .addTarget(prometheus.target('count(node_cpu_seconds_total{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster", mode="idle"})' % config, legendFormat='logical cores')), local memoryGraphPanelPrototype = graphPanel.new( 'Memory Usage', datasource='$datasource', span=9, format='bytes', min=0, ), local memoryGraph = if platform == 'Linux' then memoryGraphPanelPrototype { stack: true } .addTarget(prometheus.target( ||| ( node_memory_MemTotal_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} - node_memory_MemFree_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} - node_memory_Buffers_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} - node_memory_Cached_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} ) ||| % config, legendFormat='memory used' )) .addTarget(prometheus.target('node_memory_Buffers_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}' % config, legendFormat='memory buffers')) .addTarget(prometheus.target('node_memory_Cached_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}' % config, legendFormat='memory cached')) .addTarget(prometheus.target('node_memory_MemFree_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}' % config, legendFormat='memory free')) else if platform == 'Darwin' then // not useful to stack memoryGraphPanelPrototype { stack: false } .addTarget(prometheus.target('node_memory_total_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}' % config, legendFormat='Physical Memory')) .addTarget(prometheus.target( ||| ( node_memory_internal_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} - node_memory_purgeable_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} + node_memory_wired_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} + node_memory_compressed_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} ) ||| % config, legendFormat='Memory Used' )) .addTarget(prometheus.target( ||| ( node_memory_internal_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} - node_memory_purgeable_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} ) ||| % config, legendFormat='App Memory' )) .addTarget(prometheus.target('node_memory_wired_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}' % config, legendFormat='Wired Memory')) .addTarget(prometheus.target('node_memory_compressed_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}' % config, legendFormat='Compressed')) else if platform == 'AIX' then memoryGraphPanelPrototype { stack: false } .addTarget(prometheus.target('node_memory_total_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}' % config, legendFormat='Physical Memory')) .addTarget(prometheus.target( ||| ( node_memory_total_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} - node_memory_available_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"} ) ||| % config, legendFormat='Memory Used' )), // NOTE: avg() is used to circumvent a label change caused by a node_exporter rollout. local memoryGaugePanelPrototype = gaugePanel.new( title='Memory Usage', datasource='$datasource', ) .addThresholdStep('rgba(50, 172, 45, 0.97)') .addThresholdStep('rgba(237, 129, 40, 0.89)', 80) .addThresholdStep('rgba(245, 54, 54, 0.9)', 90) .setFieldConfig(max=100, min=0, unit='percent') + { span: 3, }, local memoryGauge = if platform == 'Linux' then memoryGaugePanelPrototype .addTarget(prometheus.target( ||| 100 - ( avg(node_memory_MemAvailable_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}) / avg(node_memory_MemTotal_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}) * 100 ) ||| % config, )) else if platform == 'Darwin' then memoryGaugePanelPrototype .addTarget(prometheus.target( ||| ( ( avg(node_memory_internal_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}) - avg(node_memory_purgeable_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}) + avg(node_memory_wired_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}) + avg(node_memory_compressed_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}) ) / avg(node_memory_total_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}) ) * 100 ||| % config )) else if platform == 'AIX' then memoryGaugePanelPrototype .addTarget(prometheus.target( ||| 100 - ( avg(node_memory_available_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}) / avg(node_memory_total_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster"}) * 100 ) ||| % config )), local diskIO = graphPanel.new( 'Disk I/O', datasource='$datasource', span=6, min=0, fill=0, ) // TODO: Does it make sense to have those three in the same panel? .addTarget(prometheus.target( 'rate(node_disk_read_bytes_total{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster", %(diskDeviceSelector)s}[$__rate_interval])' % config, legendFormat='{{device}} read', intervalFactor=1, )) .addTarget(prometheus.target( 'rate(node_disk_written_bytes_total{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster", %(diskDeviceSelector)s}[$__rate_interval])' % config, legendFormat='{{device}} written', intervalFactor=1, )) .addTarget(prometheus.target( 'rate(node_disk_io_time_seconds_total{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster", %(diskDeviceSelector)s}[$__rate_interval])' % config, legendFormat='{{device}} io time', intervalFactor=1, )) + { seriesOverrides: [ { alias: '/ read| written/', yaxis: 1, }, { alias: '/ io time/', yaxis: 2, }, ], yaxes: [ self.yaxe(format='Bps'), self.yaxe(format='percentunit'), ], }, local diskSpaceUsage = table.new( title='Disk Space Usage', datasource='$datasource', ) .setFieldConfig(unit='decbytes') .addThresholdStep(color='green', value=null) .addThresholdStep(color='yellow', value=0.8) .addThresholdStep(color='red', value=0.9) .addTarget(prometheus.target( ||| max by (mountpoint) (node_filesystem_size_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster", %(fsSelector)s, %(fsMountpointSelector)s}) ||| % config, legendFormat='', instant=true, format='table' )) .addTarget(prometheus.target( ||| max by (mountpoint) (node_filesystem_avail_bytes{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster", %(fsSelector)s, %(fsMountpointSelector)s}) ||| % config, legendFormat='', instant=true, format='table' )) .addOverride( matcher={ id: 'byName', options: 'Mounted on', }, properties=[ { id: 'custom.width', value: 260, }, ], ) .addOverride( matcher={ id: 'byName', options: 'Size', }, properties=[ { id: 'custom.width', value: 93, }, ], ) .addOverride( matcher={ id: 'byName', options: 'Used', }, properties=[ { id: 'custom.width', value: 72, }, ], ) .addOverride( matcher={ id: 'byName', options: 'Available', }, properties=[ { id: 'custom.width', value: 88, }, ], ) .addOverride( matcher={ id: 'byName', options: 'Used, %', }, properties=[ { id: 'unit', value: 'percentunit', }, { id: 'custom.displayMode', value: 'gradient-gauge', }, { id: 'max', value: 1, }, { id: 'min', value: 0, }, ] ) + { span: 6 } + { transformations: [ { id: 'groupBy', options: { fields: { 'Value #A': { aggregations: [ 'lastNotNull', ], operation: 'aggregate', }, 'Value #B': { aggregations: [ 'lastNotNull', ], operation: 'aggregate', }, mountpoint: { aggregations: [], operation: 'groupby', }, }, }, }, { id: 'merge', options: {}, }, { id: 'calculateField', options: { alias: 'Used', binary: { left: 'Value #A (lastNotNull)', operator: '-', reducer: 'sum', right: 'Value #B (lastNotNull)', }, mode: 'binary', reduce: { reducer: 'sum', }, }, }, { id: 'calculateField', options: { alias: 'Used, %', binary: { left: 'Used', operator: '/', reducer: 'sum', right: 'Value #A (lastNotNull)', }, mode: 'binary', reduce: { reducer: 'sum', }, }, }, { id: 'organize', options: { excludeByName: {}, indexByName: {}, renameByName: { 'Value #A (lastNotNull)': 'Size', 'Value #B (lastNotNull)': 'Available', mountpoint: 'Mounted on', }, }, }, { id: 'sortBy', options: { fields: {}, sort: [ { field: 'Mounted on', }, ], }, }, ], }, local networkReceived = graphPanel.new( 'Network Received', description='Network received (bits/s)', datasource='$datasource', span=6, format='bps', min=0, fill=0, ) .addTarget(prometheus.target( 'rate(node_network_receive_bytes_total{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster", device!="lo"}[$__rate_interval]) * 8' % config, legendFormat='{{device}}', intervalFactor=1, )), local networkTransmitted = graphPanel.new( 'Network Transmitted', description='Network transmitted (bits/s)', datasource='$datasource', span=6, format='bps', min=0, fill=0, ) .addTarget(prometheus.target( 'rate(node_network_transmit_bytes_total{%(nodeExporterSelector)s, instance="$instance", %(clusterLabel)s="$cluster", device!="lo"}[$__rate_interval]) * 8' % config, legendFormat='{{device}}', intervalFactor=1, )), local cpuRow = row.new('CPU') .addPanel(idleCPU) .addPanel(systemLoad), local memoryRow = row.new('Memory') .addPanel(memoryGraph) .addPanel(memoryGauge), local diskRow = row.new('Disk') .addPanel(diskIO) .addPanel(diskSpaceUsage), local networkRow = row.new('Network') .addPanel(networkReceived) .addPanel(networkTransmitted), local rows = [ cpuRow, memoryRow, diskRow, networkRow, ], local templates = [ prometheusDatasourceTemplate, clusterTemplate, instanceTemplate, ], dashboard: if platform == 'Linux' then dashboard.new( '%sNodes' % config.dashboardNamePrefix, time_from='now-1h', tags=(config.dashboardTags), timezone='utc', refresh='30s', uid=std.md5(uid), graphTooltip='shared_crosshair' ) .addTemplates(templates) .addRows(rows) else if platform == 'Darwin' then dashboard.new( '%sMacOS' % config.dashboardNamePrefix, time_from='now-1h', tags=(config.dashboardTags), timezone='utc', refresh='30s', uid=std.md5(uid), graphTooltip='shared_crosshair' ) .addTemplates(templates) .addRows(rows) else if platform == 'AIX' then dashboard.new( '%sAIX' % config.dashboardNamePrefix, time_from='now-1h', tags=(config.dashboardTags), timezone='utc', refresh='30s', uid=std.md5(uid), graphTooltip='shared_crosshair' ) .addTemplates(templates) .addRows(rows), }, }