# NOTE(srwilkers): This overrides file provides a reference for dashboards that # reflect the overall state of a Kubernetes deployment --- conf: dashboards: kubernetes_capacity_planning: |- { "__inputs": [ { "name": "DS_PROMETHEUS", "label": "prometheus", "description": "", "type": "datasource", "pluginId": "prometheus", "pluginName": "Prometheus" } ], "__requires": [ { "type": "grafana", "id": "grafana", "name": "Grafana", "version": "4.4.1" }, { "type": "panel", "id": "graph", "name": "Graph", "version": "" }, { "type": "datasource", "id": "prometheus", "name": "Prometheus", "version": "1.0.0" }, { "type": "panel", "id": "singlestat", "name": "Singlestat", "version": "" } ], "annotations": { "list": [ { "builtIn": 1, "datasource": "-- Grafana --", "enable": true, "hide": true, "iconColor": "rgba(0, 211, 255, 1)", "name": "Annotations & Alerts", "type": "dashboard" } ] }, "description": "", "overwrite": true, "editable": false, "gnetId": 22, "graphTooltip": 0, "id": 35, "links": [], "panels": [ { "alerting": {}, "aliasColors": {}, "bars": false, "dashLength": 10, "dashes": false, "datasource": "${DS_PROMETHEUS}", "editable": true, "error": false, "fill": 1, "grid": {}, "gridPos": { "h": 7, "w": 12, "x": 0, "y": 0 }, "id": 3, "legend": { "avg": false, "current": false, "max": false, "min": false, "show": true, "total": false, "values": false }, "lines": true, "linewidth": 2, "links": [], "nullPointMode": "connected", "options": {}, "percentage": false, "pointradius": 5, "points": false, "renderer": "flot", "seriesOverrides": [], "spaceLength": 10, "stack": false, "steppedLine": false, "targets": [ { "expr": "sum(rate(node_cpu{mode=\"idle\"}[2m])) * 100", "hide": false, "intervalFactor": 10, "legendFormat": "", "refId": "A", "step": 50 } ], "thresholds": [], "timeFrom": null, "timeRegions": [], "timeShift": null, "title": "Idle cpu", "tooltip": { "msResolution": false, "shared": true, "sort": 0, "value_type": "cumulative" }, "type": "graph", "xaxis": { "buckets": null, "mode": "time", "name": null, "show": true, "values": [] }, "yaxes": [ { "format": "percent", "label": "cpu usage", "logBase": 1, "max": null, "min": 0, "show": true }, { "format": "short", "label": null, "logBase": 1, "max": null, "min": null, "show": true } ], "yaxis": { "align": false, "alignLevel": null } }, { "alerting": {}, "aliasColors": {}, "bars": false, "dashLength": 10, "dashes": false, "datasource": "${DS_PROMETHEUS}", "editable": true, "error": false, "fill": 1, "grid": {}, "gridPos": { "h": 7, "w": 12, "x": 12, "y": 0 }, "id": 9, "legend": { "avg": false, "current": false, "max": false, "min": false, "show": true, "total": false, "values": false }, "lines": true, "linewidth": 2, "links": [], "nullPointMode": "connected", "options": {}, "percentage": false, "pointradius": 5, "points": false, "renderer": "flot", "seriesOverrides": [], "spaceLength": 10, "stack": false, "steppedLine": false, "targets": [ { "expr": "sum(node_load1)", "intervalFactor": 4, "legendFormat": "load 1m", "refId": "A", "step": 20, "target": "" }, { "expr": "sum(node_load5)", "intervalFactor": 4, "legendFormat": "load 5m", "refId": "B", "step": 20, "target": "" }, { "expr": "sum(node_load15)", "intervalFactor": 4, "legendFormat": "load 15m", "refId": "C", "step": 20, "target": "" } ], "thresholds": [], "timeFrom": null, "timeRegions": [], "timeShift": null, "title": "System load", "tooltip": { "msResolution": false, "shared": true, "sort": 0, "value_type": "cumulative" }, "type": "graph", "xaxis": { "buckets": null, "mode": "time", "name": null, "show": true, "values": [] }, "yaxes": [ { "format": "percentunit", "label": null, "logBase": 1, "max": null, "min": null, "show": true }, { "format": "short", "label": null, "logBase": 1, "max": null, "min": null, "show": true } ], "yaxis": { "align": false, "alignLevel": null } }, { "alerting": {}, "aliasColors": {}, "bars": false, "dashLength": 10, "dashes": false, "datasource": "${DS_PROMETHEUS}", "editable": true, "error": false, "fill": 1, "grid": {}, "gridPos": { "h": 7, "w": 18, "x": 0, "y": 7 }, "id": 4, "legend": { "avg": false, "current": false, "max": false, "min": false, "show": true, "total": false, "values": false }, "lines": true, "linewidth": 2, "links": [], "nullPointMode": "connected", "options": {}, "percentage": false, "pointradius": 5, "points": false, "renderer": "flot", "seriesOverrides": [ { "alias": "node_memory_SwapFree{instance=\"172.17.0.1:9100\",job=\"prometheus\"}", "yaxis": 2 } ], "spaceLength": 10, "stack": true, "steppedLine": false, "targets": [ { "expr": "sum(node_memory_MemTotal) - sum(node_memory_MemFree) - sum(node_memory_Buffers) - sum(node_memory_Cached)", "intervalFactor": 2, "legendFormat": "memory usage", "metric": "memo", "refId": "A", "step": 10, "target": "" }, { "expr": "sum(node_memory_Buffers)", "interval": "", "intervalFactor": 2, "legendFormat": "memory buffers", "metric": "memo", "refId": "B", "step": 10, "target": "" }, { "expr": "sum(node_memory_Cached)", "interval": "", "intervalFactor": 2, "legendFormat": "memory cached", "metric": "memo", "refId": "C", "step": 10, "target": "" }, { "expr": "sum(node_memory_MemFree)", "interval": "", "intervalFactor": 2, "legendFormat": "memory free", "metric": "memo", "refId": "D", "step": 10, "target": "" } ], "thresholds": [], "timeFrom": null, "timeRegions": [], "timeShift": null, "title": "Memory usage", "tooltip": { "msResolution": false, "shared": true, "sort": 0, "value_type": "individual" }, "type": "graph", "xaxis": { "buckets": null, "mode": "time", "name": null, "show": true, "values": [] }, "yaxes": [ { "format": "bytes", "label": null, "logBase": 1, "max": null, "min": "0", "show": true }, { "format": "short", "label": null, "logBase": 1, "max": null, "min": null, "show": true } ], "yaxis": { "align": false, "alignLevel": null } }, { "cacheTimeout": null, "colorBackground": false, "colorValue": false, "colors": [ "rgba(50, 172, 45, 0.97)", "rgba(237, 129, 40, 0.89)", "rgba(245, 54, 54, 0.9)" ], "datasource": "${DS_PROMETHEUS}", "editable": true, "error": false, "format": "percent", "gauge": { "maxValue": 100, "minValue": 0, "show": true, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 7, "w": 6, "x": 18, "y": 7 }, "id": 5, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "((sum(node_memory_MemTotal) - sum(node_memory_MemFree) - sum(node_memory_Buffers) - sum(node_memory_Cached)) / sum(node_memory_MemTotal)) * 100", "intervalFactor": 2, "metric": "", "refId": "A", "step": 60, "target": "" } ], "thresholds": "80, 90", "title": "Memory usage", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "N/A", "value": "null" } ], "valueName": "avg" }, { "alerting": {}, "aliasColors": {}, "bars": false, "dashLength": 10, "dashes": false, "datasource": "${DS_PROMETHEUS}", "editable": true, "error": false, "fill": 1, "grid": {}, "gridPos": { "h": 7, "w": 18, "x": 0, "y": 14 }, "id": 6, "legend": { "avg": false, "current": false, "max": false, "min": false, "show": true, "total": false, "values": false }, "lines": true, "linewidth": 2, "links": [], "nullPointMode": "connected", "options": {}, "percentage": false, "pointradius": 5, "points": false, "renderer": "flot", "seriesOverrides": [ { "alias": "read", "yaxis": 1 }, { "alias": "{instance=\"172.17.0.1:9100\"}", "yaxis": 2 }, { "alias": "io time", "yaxis": 2 } ], "spaceLength": 10, "stack": false, "steppedLine": false, "targets": [ { "expr": "sum(rate(node_disk_bytes_read[5m]))", "hide": false, "intervalFactor": 4, "legendFormat": "read", "refId": "A", "step": 20, "target": "" }, { "expr": "sum(rate(node_disk_bytes_written[5m]))", "intervalFactor": 4, "legendFormat": "written", "refId": "B", "step": 20 }, { "expr": "sum(rate(node_disk_io_time_ms[5m]))", "intervalFactor": 4, "legendFormat": "io time", "refId": "C", "step": 20 } ], "thresholds": [], "timeFrom": null, "timeRegions": [], "timeShift": null, "title": "Disk I/O", "tooltip": { "msResolution": false, "shared": true, "sort": 0, "value_type": "cumulative" }, "type": "graph", "xaxis": { "buckets": null, "mode": "time", "name": null, "show": true, "values": [] }, "yaxes": [ { "format": "bytes", "label": null, "logBase": 1, "max": null, "min": null, "show": true }, { "format": "ms", "label": null, "logBase": 1, "max": null, "min": null, "show": true } ], "yaxis": { "align": false, "alignLevel": null } }, { "cacheTimeout": null, "colorBackground": false, "colorValue": false, "colors": [ "rgba(50, 172, 45, 0.97)", "rgba(237, 129, 40, 0.89)", "rgba(245, 54, 54, 0.9)" ], "datasource": "${DS_PROMETHEUS}", "editable": true, "error": false, "format": "percentunit", "gauge": { "maxValue": 1, "minValue": 0, "show": true, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 7, "w": 6, "x": 18, "y": 14 }, "id": 12, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "(sum(node_filesystem_size{device!=\"rootfs\"}) - sum(node_filesystem_free{device!=\"rootfs\"})) / sum(node_filesystem_size{device!=\"rootfs\"})", "intervalFactor": 2, "refId": "A", "step": 60, "target": "" } ], "thresholds": "0.75, 0.9", "title": "Disk space usage", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "N/A", "value": "null" } ], "valueName": "current" }, { "alerting": {}, "aliasColors": {}, "bars": false, "dashLength": 10, "dashes": false, "datasource": "${DS_PROMETHEUS}", "editable": true, "error": false, "fill": 1, "grid": {}, "gridPos": { "h": 7, "w": 12, "x": 0, "y": 21 }, "id": 8, "legend": { "avg": false, "current": false, "max": false, "min": false, "show": true, "total": false, "values": false }, "lines": true, "linewidth": 2, "links": [], "nullPointMode": "connected", "options": {}, "percentage": false, "pointradius": 5, "points": false, "renderer": "flot", "seriesOverrides": [ { "alias": "transmitted ", "yaxis": 2 } ], "spaceLength": 10, "stack": false, "steppedLine": false, "targets": [ { "expr": "sum(rate(node_network_receive_bytes{device!~\"lo\"}[5m]))", "hide": false, "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 10, "target": "" } ], "thresholds": [], "timeFrom": null, "timeRegions": [], "timeShift": null, "title": "Network received", "tooltip": { "msResolution": false, "shared": true, "sort": 0, "value_type": "cumulative" }, "type": "graph", "xaxis": { "buckets": null, "mode": "time", "name": null, "show": true, "values": [] }, "yaxes": [ { "format": "bytes", "label": null, "logBase": 1, "max": null, "min": null, "show": true }, { "format": "bytes", "label": null, "logBase": 1, "max": null, "min": null, "show": true } ], "yaxis": { "align": false, "alignLevel": null } }, { "alerting": {}, "aliasColors": {}, "bars": false, "dashLength": 10, "dashes": false, "datasource": "${DS_PROMETHEUS}", "editable": true, "error": false, "fill": 1, "grid": {}, "gridPos": { "h": 7, "w": 12, "x": 12, "y": 21 }, "id": 10, "legend": { "avg": false, "current": false, "max": false, "min": false, "show": true, "total": false, "values": false }, "lines": true, "linewidth": 2, "links": [], "nullPointMode": "connected", "options": {}, "percentage": false, "pointradius": 5, "points": false, "renderer": "flot", "seriesOverrides": [ { "alias": "transmitted ", "yaxis": 2 } ], "spaceLength": 10, "stack": false, "steppedLine": false, "targets": [ { "expr": "sum(rate(node_network_transmit_bytes{device!~\"lo\"}[5m]))", "hide": false, "intervalFactor": 2, "legendFormat": "", "refId": "B", "step": 10, "target": "" } ], "thresholds": [], "timeFrom": null, "timeRegions": [], "timeShift": null, "title": "Network transmitted", "tooltip": { "msResolution": false, "shared": true, "sort": 0, "value_type": "cumulative" }, "type": "graph", "xaxis": { "buckets": null, "mode": "time", "name": null, "show": true, "values": [] }, "yaxes": [ { "format": "bytes", "label": null, "logBase": 1, "max": null, "min": null, "show": true }, { "format": "bytes", "label": null, "logBase": 1, "max": null, "min": null, "show": true } ], "yaxis": { "align": false, "alignLevel": null } }, { "aliasColors": {}, "bars": false, "dashLength": 10, "dashes": false, "datasource": "${DS_PROMETHEUS}", "fill": 1, "gridPos": { "h": 7, "w": 18, "x": 0, "y": 28 }, "id": 11, "legend": { "avg": false, "current": false, "max": false, "min": false, "show": true, "total": false, "values": false }, "lines": true, "linewidth": 1, "links": [], "nullPointMode": "null", "options": {}, "percentage": false, "pointradius": 5, "points": false, "renderer": "flot", "seriesOverrides": [], "spaceLength": 10, "stack": false, "steppedLine": false, "targets": [ { "expr": "sum(kube_pod_info)", "format": "time_series", "intervalFactor": 2, "legendFormat": "Current number of Pods", "refId": "A", "step": 10 }, { "expr": "sum(kube_node_status_capacity_pods)", "format": "time_series", "intervalFactor": 2, "legendFormat": "Maximum capacity of pods", "refId": "B", "step": 10 } ], "thresholds": [], "timeFrom": null, "timeRegions": [], "timeShift": null, "title": "Cluster Pod Utilization", "tooltip": { "shared": true, "sort": 0, "value_type": "individual" }, "type": "graph", "xaxis": { "buckets": null, "mode": "time", "name": null, "show": true, "values": [] }, "yaxes": [ { "format": "short", "label": null, "logBase": 1, "max": null, "min": null, "show": true }, { "format": "short", "label": null, "logBase": 1, "max": null, "min": null, "show": true } ], "yaxis": { "align": false, "alignLevel": null } }, { "cacheTimeout": null, "colorBackground": false, "colorValue": false, "colors": [ "rgba(50, 172, 45, 0.97)", "rgba(237, 129, 40, 0.89)", "rgba(245, 54, 54, 0.9)" ], "datasource": "${DS_PROMETHEUS}", "editable": true, "error": false, "format": "percent", "gauge": { "maxValue": 100, "minValue": 0, "show": true, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 7, "w": 6, "x": 18, "y": 28 }, "id": 7, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "100 - (sum(kube_node_status_capacity_pods) - sum(kube_pod_info)) / sum(kube_node_status_capacity_pods) * 100", "format": "time_series", "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 60, "target": "" } ], "thresholds": "80,90", "title": "Pod Utilization", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "N/A", "value": "null" } ], "valueName": "current" } ], "refresh": false, "schemaVersion": 18, "style": "dark", "tags": [], "templating": { "list": [ { "current": { "text": "prometheus", "value": "prometheus" }, "hide": 0, "includeAll": false, "label": "Prometheus datasource", "multi": false, "name": "DS_PROMETHEUS", "options": [], "query": "prometheus", "refresh": 1, "regex": "", "skipUrlSync": false, "type": "datasource" } ] }, "time": { "from": "now-1h", "to": "now" }, "timepicker": { "refresh_intervals": [ "5s", "10s", "30s", "1m", "5m", "15m", "30m", "1h", "2h", "1d" ], "time_options": [ "5m", "15m", "1h", "6h", "12h", "24h", "2d", "7d", "30d" ] }, "timezone": "browser", "title": "Kubernetes Capacity Planning", "version": 1 } kubernetes_cluster_status: |- { "__inputs": [ { "name": "DS_PROMETHEUS", "label": "prometheus", "description": "", "type": "datasource", "pluginId": "prometheus", "pluginName": "Prometheus" } ], "__requires": [ { "type": "grafana", "id": "grafana", "name": "Grafana", "version": "4.4.1" }, { "type": "panel", "id": "graph", "name": "Graph", "version": "" }, { "type": "datasource", "id": "prometheus", "name": "Prometheus", "version": "1.0.0" }, { "type": "panel", "id": "singlestat", "name": "Singlestat", "version": "" } ], "annotations": { "list": [ { "builtIn": 1, "datasource": "-- Grafana --", "enable": true, "hide": true, "iconColor": "rgba(0, 211, 255, 1)", "name": "Annotations & Alerts", "type": "dashboard" } ] }, "editable": false, "overwrite": true, "gnetId": null, "graphTooltip": 0, "id": 5, "links": [], "panels": [ { "collapsed": false, "gridPos": { "h": 1, "w": 24, "x": 0, "y": 0 }, "id": 11, "panels": [], "repeat": null, "title": "Cluster Health", "type": "row" }, { "cacheTimeout": null, "colorBackground": false, "colorValue": true, "colors": [ "rgba(50, 172, 45, 0.97)", "rgba(237, 129, 40, 0.89)", "rgba(245, 54, 54, 0.9)" ], "datasource": "${DS_PROMETHEUS}", "format": "none", "gauge": { "maxValue": 100, "minValue": 0, "show": false, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 4, "w": 12, "x": 0, "y": 1 }, "id": 5, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "sum(up{job=~\"apiserver|kube-scheduler|kube-controller-manager\"} == 0)", "format": "time_series", "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 600 } ], "thresholds": "1,3", "title": "Control Plane UP", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "UP", "value": "null" } ], "valueName": "total" }, { "cacheTimeout": null, "colorBackground": false, "colorValue": true, "colors": [ "rgba(50, 172, 45, 0.97)", "rgba(237, 129, 40, 0.89)", "rgba(245, 54, 54, 0.9)" ], "datasource": "${DS_PROMETHEUS}", "format": "none", "gauge": { "maxValue": 100, "minValue": 0, "show": false, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 4, "w": 12, "x": 12, "y": 1 }, "id": 6, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "sum(ALERTS{alertstate=\"firing\",alertname!=\"DeadMansSwitch\"})", "format": "time_series", "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 600 } ], "thresholds": "3,5", "title": "Alerts Firing", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "0", "value": "null" } ], "valueName": "current" }, { "collapsed": false, "gridPos": { "h": 1, "w": 24, "x": 0, "y": 5 }, "id": 12, "panels": [], "repeat": null, "title": "Control Plane Status", "type": "row" }, { "cacheTimeout": null, "colorBackground": false, "colorValue": false, "colors": [ "rgba(245, 54, 54, 0.9)", "rgba(237, 129, 40, 0.89)", "rgba(50, 172, 45, 0.97)" ], "datasource": "${DS_PROMETHEUS}", "decimals": null, "format": "percent", "gauge": { "maxValue": 100, "minValue": 0, "show": true, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 5, "w": 6, "x": 0, "y": 6 }, "id": 1, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "(sum(up{job=\"apiserver\"} == 1) / count(up{job=\"apiserver\"})) * 100", "format": "time_series", "interval": "", "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 600 } ], "thresholds": "50,80", "title": "API Servers UP", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "N/A", "value": "null" } ], "valueName": "current" }, { "cacheTimeout": null, "colorBackground": false, "colorValue": false, "colors": [ "rgba(245, 54, 54, 0.9)", "rgba(237, 129, 40, 0.89)", "rgba(50, 172, 45, 0.97)" ], "datasource": "${DS_PROMETHEUS}", "decimals": null, "format": "percent", "gauge": { "maxValue": 100, "minValue": 0, "show": true, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 5, "w": 6, "x": 6, "y": 6 }, "id": 2, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "(sum(up{job=\"kube-controller-manager-discovery\"} == 1) / count(up{job=\"kube-controller-manager-discovery\"})) * 100", "format": "time_series", "interval": "", "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 600 } ], "thresholds": "50,80", "title": "Controller Managers UP", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "N/A", "value": "null" } ], "valueName": "current" }, { "cacheTimeout": null, "colorBackground": false, "colorValue": false, "colors": [ "rgba(245, 54, 54, 0.9)", "rgba(237, 129, 40, 0.89)", "rgba(50, 172, 45, 0.97)" ], "datasource": "${DS_PROMETHEUS}", "decimals": null, "format": "percent", "gauge": { "maxValue": 100, "minValue": 0, "show": true, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 5, "w": 6, "x": 12, "y": 6 }, "id": 3, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "(sum(up{job=\"kube-scheduler-discovery\"} == 1) / count(up{job=\"kube-scheduler-discovery\"})) * 100", "format": "time_series", "interval": "", "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 600 } ], "thresholds": "50,80", "title": "Schedulers UP", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "N/A", "value": "null" } ], "valueName": "current" }, { "cacheTimeout": null, "colorBackground": false, "colorValue": true, "colors": [ "rgba(50, 172, 45, 0.97)", "rgba(237, 129, 40, 0.89)", "rgba(245, 54, 54, 0.9)" ], "datasource": "${DS_PROMETHEUS}", "decimals": null, "format": "none", "gauge": { "maxValue": 100, "minValue": 0, "show": false, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 5, "w": 6, "x": 18, "y": 6 }, "hideTimeOverride": false, "id": 4, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "count(increase(kube_pod_container_status_restarts{namespace=~\"kube-system|tectonic-system\"}[1h]) > 5)", "format": "time_series", "interval": "", "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 600 } ], "thresholds": "1,3", "title": "Crashlooping Control Plane Pods", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "0", "value": "null" } ], "valueName": "current" }, { "collapsed": false, "gridPos": { "h": 1, "w": 24, "x": 0, "y": 11 }, "id": 13, "panels": [], "repeat": null, "title": "Capacity Planing", "type": "row" }, { "cacheTimeout": null, "colorBackground": false, "colorValue": false, "colors": [ "rgba(50, 172, 45, 0.97)", "rgba(237, 129, 40, 0.89)", "rgba(245, 54, 54, 0.9)" ], "datasource": "${DS_PROMETHEUS}", "format": "percent", "gauge": { "maxValue": 100, "minValue": 0, "show": true, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 4, "w": 6, "x": 0, "y": 12 }, "id": 8, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "sum(100 - (avg by (instance) (rate(node_cpu{job=\"node-exporter\",mode=\"idle\"}[5m])) * 100)) / count(node_cpu{job=\"node-exporter\",mode=\"idle\"})", "format": "time_series", "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 600 } ], "thresholds": "80,90", "title": "CPU Utilization", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "N/A", "value": "null" } ], "valueName": "avg" }, { "cacheTimeout": null, "colorBackground": false, "colorValue": false, "colors": [ "rgba(50, 172, 45, 0.97)", "rgba(237, 129, 40, 0.89)", "rgba(245, 54, 54, 0.9)" ], "datasource": "${DS_PROMETHEUS}", "format": "percent", "gauge": { "maxValue": 100, "minValue": 0, "show": true, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 4, "w": 6, "x": 6, "y": 12 }, "id": 7, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "((sum(node_memory_MemTotal) - sum(node_memory_MemFree) - sum(node_memory_Buffers) - sum(node_memory_Cached)) / sum(node_memory_MemTotal)) * 100", "format": "time_series", "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 600 } ], "thresholds": "80,90", "title": "Memory Utilization", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "N/A", "value": "null" } ], "valueName": "avg" }, { "cacheTimeout": null, "colorBackground": false, "colorValue": false, "colors": [ "rgba(50, 172, 45, 0.97)", "rgba(237, 129, 40, 0.89)", "rgba(245, 54, 54, 0.9)" ], "datasource": "${DS_PROMETHEUS}", "format": "percent", "gauge": { "maxValue": 100, "minValue": 0, "show": true, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 4, "w": 6, "x": 12, "y": 12 }, "id": 9, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "(sum(node_filesystem_size{device!=\"rootfs\"}) - sum(node_filesystem_free{device!=\"rootfs\"})) / sum(node_filesystem_size{device!=\"rootfs\"})", "format": "time_series", "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 600 } ], "thresholds": "80,90", "title": "Filesystem Utilization", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "N/A", "value": "null" } ], "valueName": "avg" }, { "cacheTimeout": null, "colorBackground": false, "colorValue": false, "colors": [ "rgba(50, 172, 45, 0.97)", "rgba(237, 129, 40, 0.89)", "rgba(245, 54, 54, 0.9)" ], "datasource": "${DS_PROMETHEUS}", "format": "percent", "gauge": { "maxValue": 100, "minValue": 0, "show": true, "thresholdLabels": false, "thresholdMarkers": true }, "gridPos": { "h": 4, "w": 6, "x": 18, "y": 12 }, "id": 10, "interval": null, "links": [], "mappingType": 1, "mappingTypes": [ { "name": "value to text", "value": 1 }, { "name": "range to text", "value": 2 } ], "maxDataPoints": 100, "nullPointMode": "connected", "nullText": null, "options": {}, "postfix": "", "postfixFontSize": "50%", "prefix": "", "prefixFontSize": "50%", "rangeMaps": [ { "from": "null", "text": "N/A", "to": "null" } ], "sparkline": { "fillColor": "rgba(31, 118, 189, 0.18)", "full": false, "lineColor": "rgb(31, 120, 193)", "show": false }, "tableColumn": "", "targets": [ { "expr": "100 - (sum(kube_node_status_capacity_pods) - sum(kube_pod_info)) / sum(kube_node_status_capacity_pods) * 100", "format": "time_series", "intervalFactor": 2, "legendFormat": "", "refId": "A", "step": 600 } ], "thresholds": "80,90", "title": "Pod Utilization", "type": "singlestat", "valueFontSize": "80%", "valueMaps": [ { "op": "=", "text": "N/A", "value": "null" } ], "valueName": "avg" } ], "schemaVersion": 18, "style": "dark", "tags": [], "templating": { "list": [ { "current": { "text": "prometheus", "value": "prometheus" }, "hide": 0, "includeAll": false, "label": "Prometheus datasource", "multi": false, "name": "DS_PROMETHEUS", "options": [], "query": "prometheus", "refresh": 1, "regex": "", "skipUrlSync": false, "type": "datasource" } ] }, "time": { "from": "now-1h", "to": "now" }, "timepicker": { "refresh_intervals": [ "5s", "10s", "30s", "1m", "5m", "15m", "30m", "1h", "2h", "1d" ], "time_options": [ "5m", "15m", "1h", "6h", "12h", "24h", "2d", "7d", "30d" ] }, "timezone": "browser", "title": "Kubernetes Cluster Status", "version": 1 } ...