infra/modules/kubernetes/monitoring/dashboards/cluster_health.json

5000 lines
113 KiB
JSON

{
"annotations": {
"list": [
{
"builtIn": 1,
"datasource": {
"type": "datasource",
"uid": "grafana"
},
"enable": true,
"hide": true,
"iconColor": "rgba(0, 211, 255, 1)",
"name": "Annotations & Alerts",
"type": "dashboard"
}
]
},
"description": "Single-pane-of-glass daily health overview — nodes, pods, quotas, storage, certs, GPU, power, smart home",
"editable": true,
"fiscalYearStartMonth": 0,
"graphTooltip": 1,
"id": 0,
"links": [
{
"asDropdown": true,
"icon": "external link",
"includeVars": false,
"keepTime": true,
"tags": [],
"targetBlank": true,
"title": "Dashboards",
"type": "dashboards",
"url": ""
},
{
"icon": "cloud",
"keepTime": true,
"targetBlank": true,
"title": "API Server",
"tooltip": "Kubernetes API Server",
"type": "link",
"url": "/d/k8s_system_apisrv"
},
{
"icon": "dashboard",
"keepTime": true,
"targetBlank": true,
"title": "Nodes",
"tooltip": "Kubernetes Nodes View",
"type": "link",
"url": "/d/k8s_views_nodes"
},
{
"icon": "dashboard",
"keepTime": true,
"targetBlank": true,
"title": "Pods",
"tooltip": "Kubernetes Pods View",
"type": "link",
"url": "/d/k8s_views_pods"
},
{
"icon": "bolt",
"keepTime": true,
"targetBlank": true,
"title": "GPU",
"tooltip": "NVIDIA DCGM Exporter",
"type": "link",
"url": "/d/Oxed_c6Wz"
},
{
"icon": "server",
"keepTime": true,
"targetBlank": true,
"title": "iDRAC",
"tooltip": "Dell iDRAC Hardware",
"type": "link",
"url": "/d/YVz226S4z"
},
{
"icon": "battery-full",
"keepTime": true,
"targetBlank": true,
"title": "UPS",
"tooltip": "Huawei UPS 2000",
"type": "link",
"url": "/d/ee70yskqw5u68f"
},
{
"icon": "signal",
"keepTime": true,
"targetBlank": true,
"title": "CoreDNS",
"tooltip": "CoreDNS",
"type": "link",
"url": "/d/wY4blRMGz"
},
{
"icon": "monitor",
"keepTime": true,
"targetBlank": true,
"title": "Node Exporter",
"tooltip": "Node Exporter Full",
"type": "link",
"url": "/d/rYdddlPWk"
},
{
"icon": "cube",
"keepTime": true,
"targetBlank": true,
"title": "Docker Registry",
"tooltip": "Docker Registry",
"type": "link",
"url": "/d/CoBSgj8iz"
},
{
"icon": "exchange-alt",
"keepTime": true,
"targetBlank": true,
"title": "Traefik",
"tooltip": "Traefik Ingress",
"type": "link",
"url": "/d/n5bu_kv45"
},
{
"icon": "file-alt",
"keepTime": true,
"targetBlank": true,
"title": "Loki Logs",
"tooltip": "Loki Kubernetes Logs",
"type": "link",
"url": "/d/o6-BGgnnk"
},
{
"icon": "graph-bar",
"keepTime": true,
"targetBlank": true,
"title": "kube-state-metrics",
"tooltip": "kube-state-metrics v2",
"type": "link",
"url": "/d/garysdevil-kube-state-metrics-v2"
},
{
"icon": "server",
"keepTime": true,
"targetBlank": true,
"title": "Proxmox",
"tooltip": "Proxmox Node Exporter",
"type": "link",
"url": "/d/rYdddlPW"
}
],
"panels": [
{
"collapsed": false,
"gridPos": {
"h": 1,
"w": 24,
"x": 0,
"y": 0
},
"id": 70,
"panels": [],
"title": "Infrastructure",
"type": "row"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "orange",
"value": 70
},
{
"color": "red",
"value": 85
}
]
},
"unit": "celsius"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 0,
"y": 1
},
"id": 78,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "auto",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "max(r730_idrac_idrac_sensors_temperature{name=\"CPU1 Temp\"})",
"legendFormat": "CPU Temp",
"refId": "A"
}
],
"title": "CPU Temp",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "orange",
"value": 70
},
{
"color": "red",
"value": 85
}
]
},
"unit": "celsius"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 3,
"y": 1
},
"id": 71,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "auto",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "max(nvidia_tesla_t4_DCGM_FI_DEV_GPU_TEMP)",
"legendFormat": "GPU Temp",
"refId": "A"
}
],
"title": "GPU Temp",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [
{
"options": {
"from": 1,
"result": {
"color": "red",
"text": "OUTAGE"
},
"to": 999999
},
"type": "range"
},
{
"options": {
"0": {
"color": "green",
"text": "MAINS"
}
},
"type": "value"
}
],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "red",
"value": 1
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 6,
"y": 1
},
"id": 93,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "auto",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "max(ups_upsSecondsOnBattery)",
"legendFormat": "",
"refId": "A"
}
],
"title": "Power Source",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "orange",
"value": 60
},
{
"color": "red",
"value": 85
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 9,
"y": 1
},
"id": 72,
"options": {
"minVizHeight": 75,
"minVizWidth": 75,
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showThresholdLabels": false,
"showThresholdMarkers": true,
"sizing": "auto"
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "max(nvidia_tesla_t4_DCGM_FI_DEV_GPU_UTIL)",
"legendFormat": "GPU Util",
"refId": "A"
}
],
"title": "GPU Util %",
"type": "gauge"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "orange",
"value": 12000
},
{
"color": "red",
"value": 14000
}
]
},
"unit": "decmbytes"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 12,
"y": 1
},
"id": 73,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "auto",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "max(nvidia_tesla_t4_DCGM_FI_DEV_FB_USED)",
"legendFormat": "VRAM",
"refId": "A"
}
],
"title": "GPU VRAM Used",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "red",
"value": 0
},
{
"color": "orange",
"value": 10
},
{
"color": "green",
"value": 30
}
]
},
"unit": "m"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 15,
"y": 1
},
"id": 75,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "auto",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "max(ups_upsEstimatedMinutesRemaining)",
"legendFormat": "Battery",
"refId": "A"
}
],
"title": "UPS Battery (min)",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "red",
"value": 0
},
{
"color": "orange",
"value": 30
},
{
"color": "green",
"value": 80
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 18,
"y": 1
},
"id": 76,
"options": {
"minVizHeight": 75,
"minVizWidth": 75,
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showThresholdLabels": false,
"showThresholdMarkers": true,
"sizing": "auto"
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "max(ups_upsEstimatedChargeRemaining)",
"legendFormat": "Charge",
"refId": "A"
}
],
"title": "UPS Charge %",
"type": "gauge"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [
{
"options": {
"0": {
"color": "green",
"text": "0"
}
},
"type": "value"
}
],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "red",
"value": 1
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 21,
"y": 1
},
"id": 94,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "auto",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "max(ups_upsAlarmsPresent)",
"legendFormat": "",
"refId": "A"
}
],
"title": "UPS Alarms",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisBorderShow": false,
"axisCenteredZero": false,
"axisColorMode": "text",
"axisLabel": "",
"axisPlacement": "auto",
"barAlignment": 0,
"barWidthFactor": 0.6,
"drawStyle": "line",
"fillOpacity": 10,
"gradientMode": "none",
"hideFrom": {
"legend": false,
"tooltip": false,
"viz": false
},
"insertNulls": false,
"lineInterpolation": "smooth",
"lineWidth": 2,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"showValues": false,
"spanNulls": true,
"stacking": {
"group": "A",
"mode": "none"
},
"thresholdsStyle": {
"mode": "off"
}
},
"mappings": [],
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
}
]
},
"unit": "watt"
},
"overrides": [
{
"matcher": {
"id": "byName",
"options": "Solar Production"
},
"properties": [
{
"id": "color",
"value": {
"fixedColor": "#FADE2A",
"mode": "fixed"
}
},
{
"id": "custom.fillOpacity",
"value": 20
}
]
},
{
"matcher": {
"id": "byName",
"options": "CPU Temp"
},
"properties": [
{
"id": "custom.axisPlacement",
"value": "right"
},
{
"id": "unit",
"value": "celsius"
},
{
"id": "color",
"value": {
"fixedColor": "#FF9830",
"mode": "fixed"
}
},
{
"id": "custom.fillOpacity",
"value": 0
}
]
}
]
},
"gridPos": {
"h": 8,
"w": 24,
"x": 0,
"y": 5
},
"id": 77,
"options": {
"legend": {
"calcs": [
"mean",
"max",
"min"
],
"displayMode": "table",
"placement": "bottom",
"showLegend": true
},
"tooltip": {
"hideZeros": false,
"mode": "multi",
"sort": "desc"
}
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "max by() (avg_over_time(r730_idrac_idrac_power_control_avg_consumed_watts[$__rate_interval]))",
"legendFormat": "Consumed",
"refId": "A"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "max(haos_sensor_power_w{entity=\"sensor.fv_b_pv_power\"})",
"legendFormat": "Solar Production",
"refId": "B"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "max(r730_idrac_idrac_sensors_temperature{name=\"CPU1 Temp\"})",
"legendFormat": "CPU Temp",
"refId": "C"
}
],
"title": "Power Draw (24h)",
"type": "timeseries"
},
{
"collapsed": false,
"gridPos": {
"h": 1,
"w": 24,
"x": 0,
"y": 13
},
"id": 100,
"panels": [],
"title": "Smart Home (ha-sofia)",
"type": "row"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [
{
"options": {
"0": {
"color": "red",
"text": "DOWN"
},
"1": {
"color": "green",
"text": "UP"
}
},
"type": "value"
}
],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "red",
"value": 0
},
{
"color": "green",
"value": 1
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 0,
"y": 14
},
"id": 101,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "auto",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "up{job=\"haos\"}",
"legendFormat": "",
"refId": "A"
}
],
"title": "HA Status",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "red",
"value": 0
},
{
"color": "orange",
"value": 70
},
{
"color": "green",
"value": 90
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 3,
"y": 14
},
"id": 102,
"options": {
"minVizHeight": 75,
"minVizWidth": 75,
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showThresholdLabels": false,
"showThresholdMarkers": true,
"sizing": "auto"
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=~\"alarm_control_panel|camera|climate|fan|humidifier|light|media_player|remote|switch|vacuum\"} == 1) / count(haos_entity_available{domain=~\"alarm_control_panel|camera|climate|fan|humidifier|light|media_player|remote|switch|vacuum\"}) * 100",
"legendFormat": "",
"refId": "A"
}
],
"title": "Devices Available",
"type": "gauge"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "red",
"value": 1
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 7,
"y": 14
},
"id": 103,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "auto",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=~\"alarm_control_panel|camera|climate|fan|humidifier|light|media_player|remote|switch|vacuum\"} == 0) OR vector(0)",
"legendFormat": "",
"refId": "A"
}
],
"title": "Unavailable Devices",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "orange",
"value": 0.01
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 2,
"x": 10,
"y": 14
},
"id": 104,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "value_and_name",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"light\"} == 1)",
"legendFormat": "avail",
"refId": "A"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"light\"})",
"legendFormat": "total",
"refId": "B"
}
],
"title": "Lights",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "orange",
"value": 0.01
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 2,
"x": 12,
"y": 14
},
"id": 105,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "value_and_name",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"camera\"} == 1)",
"legendFormat": "avail",
"refId": "A"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"camera\"})",
"legendFormat": "total",
"refId": "B"
}
],
"title": "Cameras",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "orange",
"value": 0.01
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 2,
"x": 14,
"y": 14
},
"id": 106,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "value_and_name",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"climate\"} == 1)",
"legendFormat": "avail",
"refId": "A"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"climate\"})",
"legendFormat": "total",
"refId": "B"
}
],
"title": "Climate",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "orange",
"value": 0.01
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 16,
"y": 14
},
"id": 107,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "value_and_name",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"switch\"} == 1)",
"legendFormat": "avail",
"refId": "A"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"switch\"})",
"legendFormat": "total",
"refId": "B"
}
],
"title": "Switches",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "orange",
"value": 0.01
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 2,
"x": 19,
"y": 14
},
"id": 108,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "value_and_name",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"alarm_control_panel\"} == 1)",
"legendFormat": "avail",
"refId": "A"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"alarm_control_panel\"})",
"legendFormat": "total",
"refId": "B"
}
],
"title": "Alarm",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "orange",
"value": 0.01
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 21,
"y": 14
},
"id": 109,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"percentChangeColorMode": "standard",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showPercentChange": false,
"textMode": "value_and_name",
"wideLayout": true
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"media_player\"} == 1)",
"legendFormat": "avail",
"refId": "A"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=\"media_player\"})",
"legendFormat": "total",
"refId": "B"
}
],
"title": "Media Players",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"custom": {
"align": "auto",
"cellOptions": {
"type": "auto"
},
"footer": {
"reducers": []
},
"inspect": false
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "red",
"value": 0
}
]
}
},
"overrides": [
{
"matcher": {
"id": "byName",
"options": "Domain"
},
"properties": [
{
"id": "custom.width",
"value": 180
}
]
},
{
"matcher": {
"id": "byName",
"options": "Name"
},
"properties": [
{
"id": "custom.width",
"value": 350
}
]
}
]
},
"gridPos": {
"h": 8,
"w": 14,
"x": 0,
"y": 18
},
"id": 110,
"options": {
"cellHeight": "sm",
"showHeader": true,
"sortBy": [
{
"desc": false,
"displayName": "Domain"
}
]
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"editorMode": "code",
"expr": "haos_entity_available{domain=~\"alarm_control_panel|camera|climate|fan|humidifier|light|media_player|remote|switch|vacuum\"} == 0",
"format": "table",
"instant": true,
"legendFormat": "",
"refId": "A"
}
],
"title": "Unavailable Devices",
"transformations": [
{
"id": "organize",
"options": {
"excludeByName": {
"Time": true,
"Value": true,
"__name__": true,
"instance": true,
"job": true
},
"renameByName": {
"domain": "Domain",
"entity": "Entity",
"friendly_name": "Name"
}
}
},
{
"id": "sortBy",
"options": {
"fields": {},
"sort": [
{
"desc": false,
"field": "Domain"
}
]
}
}
],
"type": "table"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisBorderShow": false,
"axisCenteredZero": false,
"axisColorMode": "text",
"axisLabel": "",
"axisPlacement": "auto",
"barAlignment": 0,
"barWidthFactor": 0.6,
"drawStyle": "line",
"fillOpacity": 20,
"gradientMode": "scheme",
"hideFrom": {
"legend": false,
"tooltip": false,
"viz": false
},
"insertNulls": false,
"lineInterpolation": "smooth",
"lineWidth": 2,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"showValues": false,
"spanNulls": true,
"stacking": {
"group": "A",
"mode": "none"
},
"thresholdsStyle": {
"mode": "off"
}
},
"mappings": [],
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
}
]
}
},
"overrides": [
{
"matcher": {
"id": "byName",
"options": "Unavailable"
},
"properties": [
{
"id": "color",
"value": {
"fixedColor": "red",
"mode": "fixed"
}
}
]
},
{
"matcher": {
"id": "byName",
"options": "Available"
},
"properties": [
{
"id": "color",
"value": {
"fixedColor": "green",
"mode": "fixed"
}
}
]
}
]
},
"gridPos": {
"h": 8,
"w": 10,
"x": 14,
"y": 18
},
"id": 111,
"options": {
"legend": {
"calcs": [
"lastNotNull",
"min",
"max"
],
"displayMode": "table",
"placement": "bottom",
"showLegend": true
},
"tooltip": {
"hideZeros": false,
"mode": "multi",
"sort": "none"
}
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=~\"alarm_control_panel|camera|climate|fan|humidifier|light|media_player|remote|switch|vacuum\"} == 1)",
"legendFormat": "Available",
"refId": "A"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(haos_entity_available{domain=~\"alarm_control_panel|camera|climate|fan|humidifier|light|media_player|remote|switch|vacuum\"} == 0)",
"legendFormat": "Unavailable",
"refId": "B"
}
],
"title": "Device Availability Over Time",
"type": "timeseries"
},
{
"collapsed": false,
"gridPos": {
"h": 1,
"w": 24,
"x": 0,
"y": 26
},
"id": 10,
"panels": [],
"title": "Node Health",
"type": "row"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"custom": {
"align": "auto",
"cellOptions": {
"type": "auto"
},
"footer": {
"reducers": []
},
"inspect": false
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "orange",
"value": 60
},
{
"color": "red",
"value": 80
}
]
}
},
"overrides": [
{
"matcher": {
"id": "byName",
"options": "Node"
},
"properties": [
{
"id": "custom.width",
"value": 160
}
]
},
{
"matcher": {
"id": "byName",
"options": "CPU %"
},
"properties": [
{
"id": "unit",
"value": "percent"
},
{
"id": "decimals",
"value": 1
},
{
"id": "custom.cellOptions",
"value": {
"mode": "gradient",
"type": "gauge"
}
}
]
},
{
"matcher": {
"id": "byName",
"options": "Memory %"
},
"properties": [
{
"id": "unit",
"value": "percent"
},
{
"id": "decimals",
"value": 1
},
{
"id": "custom.cellOptions",
"value": {
"mode": "gradient",
"type": "gauge"
}
}
]
},
{
"matcher": {
"id": "byName",
"options": "Disk %"
},
"properties": [
{
"id": "unit",
"value": "percent"
},
{
"id": "decimals",
"value": 1
},
{
"id": "custom.cellOptions",
"value": {
"mode": "gradient",
"type": "gauge"
}
}
]
},
{
"matcher": {
"id": "byName",
"options": "Pods"
},
"properties": [
{
"id": "custom.width",
"value": 60
}
]
}
]
},
"gridPos": {
"h": 8,
"w": 8,
"x": 0,
"y": 27
},
"id": 11,
"options": {
"cellHeight": "sm",
"showHeader": true,
"sortBy": [
{
"desc": true,
"displayName": "CPU %"
}
]
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "(1 - avg by(instance)(rate(node_cpu_seconds_total{mode=\"idle\"}[5m]))) * 100",
"format": "table",
"instant": true,
"legendFormat": "",
"refId": "CPU"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "(1 - node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes) * 100",
"format": "table",
"instant": true,
"legendFormat": "",
"refId": "MEM"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "(1 - node_filesystem_avail_bytes{mountpoint=\"/\",fstype!~\"tmpfs\"} / node_filesystem_size_bytes{mountpoint=\"/\",fstype!~\"tmpfs\"}) * 100",
"format": "table",
"instant": true,
"legendFormat": "",
"refId": "DISK"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count by(node)(kube_pod_info)",
"format": "table",
"instant": true,
"legendFormat": "",
"refId": "PODS"
}
],
"title": "Node Resource Table",
"transformations": [
{
"id": "merge",
"options": {}
},
{
"id": "organize",
"options": {
"excludeByName": {
"Time": true
},
"renameByName": {
"Value #CPU": "CPU %",
"Value #DISK": "Disk %",
"Value #MEM": "Memory %",
"Value #PODS": "Pods",
"instance": "Node",
"node": "Node"
}
}
}
],
"type": "table"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisBorderShow": false,
"axisCenteredZero": false,
"axisColorMode": "text",
"axisLabel": "",
"axisPlacement": "auto",
"barAlignment": 0,
"barWidthFactor": 0.6,
"drawStyle": "line",
"fillOpacity": 10,
"gradientMode": "none",
"hideFrom": {
"legend": false,
"tooltip": false,
"viz": false
},
"insertNulls": false,
"lineInterpolation": "smooth",
"lineWidth": 2,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"showValues": false,
"spanNulls": false,
"stacking": {
"group": "A",
"mode": "none"
},
"thresholdsStyle": {
"mode": "off"
}
},
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "red",
"value": 80
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 8,
"w": 8,
"x": 8,
"y": 27
},
"id": 12,
"options": {
"legend": {
"calcs": [
"mean",
"max"
],
"displayMode": "table",
"placement": "bottom",
"showLegend": true
},
"tooltip": {
"hideZeros": false,
"mode": "multi",
"sort": "desc"
}
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "(1 - avg by(instance)(rate(node_cpu_seconds_total{mode=\"idle\"}[5m]))) * 100",
"legendFormat": "{{instance}}",
"refId": "A"
}
],
"title": "CPU per Node (24h)",
"type": "timeseries"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisBorderShow": false,
"axisCenteredZero": false,
"axisColorMode": "text",
"axisLabel": "",
"axisPlacement": "auto",
"barAlignment": 0,
"barWidthFactor": 0.6,
"drawStyle": "line",
"fillOpacity": 10,
"gradientMode": "none",
"hideFrom": {
"legend": false,
"tooltip": false,
"viz": false
},
"insertNulls": false,
"lineInterpolation": "smooth",
"lineWidth": 2,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"showValues": false,
"spanNulls": false,
"stacking": {
"group": "A",
"mode": "none"
},
"thresholdsStyle": {
"mode": "off"
}
},
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green",
"value": 0
},
{
"color": "red",
"value": 80
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 8,
"w": 8,
"x": 16,
"y": 27
},
"id": 13,
"options": {
"legend": {
"calcs": [
"mean",
"max"
],
"displayMode": "table",
"placement": "bottom",
"showLegend": true
},
"tooltip": {
"hideZeros": false,
"mode": "multi",
"sort": "desc"
}
},
"pluginVersion": "12.3.1",
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "(1 - node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes) * 100",
"legendFormat": "{{instance}}",
"refId": "A"
}
],
"title": "Memory per Node (24h)",
"type": "timeseries"
},
{
"collapsed": false,
"gridPos": {
"h": 1,
"w": 24,
"x": 0,
"y": 35
},
"id": 40,
"panels": [],
"title": "Top Resource Consumers",
"type": "row"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
}
]
},
"unit": "short"
},
"overrides": []
},
"gridPos": {
"h": 8,
"w": 8,
"x": 0,
"y": 36
},
"id": 41,
"options": {
"displayMode": "gradient",
"maxVizHeight": 300,
"minVizHeight": 16,
"minVizWidth": 8,
"namePlacement": "auto",
"orientation": "horizontal",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showUnfilled": true,
"sizing": "auto",
"valueMode": "color"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "topk(10, sum by(namespace)(rate(container_cpu_usage_seconds_total{container!=\"\"}[5m])))",
"format": "table",
"instant": true,
"refId": "A"
}
],
"title": "Top 10 Namespaces by CPU",
"transformations": [
{
"id": "organize",
"options": {
"excludeByName": {
"Time": true,
"__name__": true,
"container": true,
"endpoint": true,
"id": true,
"instance": true,
"job": true,
"metrics_path": true,
"node": true,
"service": true,
"uid": true
},
"renameByName": {}
}
},
{
"id": "sortBy",
"options": {
"fields": {},
"sort": [
{
"desc": true,
"field": "Value"
}
]
}
},
{
"id": "rowsToFields",
"options": {}
}
],
"type": "bargauge"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
}
]
},
"unit": "bytes"
},
"overrides": []
},
"gridPos": {
"h": 8,
"w": 8,
"x": 8,
"y": 36
},
"id": 42,
"options": {
"displayMode": "gradient",
"maxVizHeight": 300,
"minVizHeight": 16,
"minVizWidth": 8,
"namePlacement": "auto",
"orientation": "horizontal",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showUnfilled": true,
"sizing": "auto",
"valueMode": "color"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "topk(10, sum by(namespace)(container_memory_working_set_bytes{container!=\"\"}))",
"format": "table",
"instant": true,
"refId": "A"
}
],
"title": "Top 10 Namespaces by Memory",
"transformations": [
{
"id": "organize",
"options": {
"excludeByName": {
"Time": true,
"__name__": true,
"container": true,
"endpoint": true,
"id": true,
"instance": true,
"job": true,
"metrics_path": true,
"node": true,
"service": true,
"uid": true
},
"renameByName": {}
}
},
{
"id": "sortBy",
"options": {
"fields": {},
"sort": [
{
"desc": true,
"field": "Value"
}
]
}
},
{
"id": "rowsToFields",
"options": {}
}
],
"type": "bargauge"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisBorderShow": false,
"axisCenteredZero": false,
"axisColorMode": "text",
"axisLabel": "",
"axisPlacement": "auto",
"barAlignment": 0,
"barWidthFactor": 0.6,
"drawStyle": "line",
"fillOpacity": 20,
"gradientMode": "none",
"hideFrom": {
"legend": false,
"tooltip": false,
"viz": false
},
"insertNulls": false,
"lineInterpolation": "smooth",
"lineWidth": 2,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"spanNulls": true,
"stacking": {
"group": "A",
"mode": "normal"
},
"thresholdsStyle": {
"mode": "off"
}
},
"mappings": [],
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
}
]
},
"unit": "bytes"
},
"overrides": []
},
"gridPos": {
"h": 8,
"w": 8,
"x": 16,
"y": 36
},
"id": 43,
"options": {
"legend": {
"calcs": [
"lastNotNull"
],
"displayMode": "table",
"placement": "bottom",
"showLegend": true
},
"tooltip": {
"mode": "multi",
"sort": "desc"
}
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "sum by (namespace) (gpu_pod_memory_used_bytes) > 0",
"format": "time_series",
"instant": false,
"legendFormat": "{{namespace}}",
"refId": "A"
}
],
"title": "GPU Memory by Namespace",
"type": "timeseries"
},
{
"collapsed": false,
"gridPos": {
"h": 1,
"w": 24,
"x": 0,
"y": 44
},
"id": 80,
"panels": [],
"title": "Key Services",
"type": "row"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"decimals": 0,
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "blue"
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 0,
"y": 45
},
"id": 87,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "sum(max by(entrypoint, protocol) (traefik_open_connections))",
"legendFormat": "",
"refId": "A"
}
],
"title": "Open Connections",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"decimals": 1,
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "red"
},
{
"color": "orange",
"value": 50
},
{
"color": "green",
"value": 80
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 8,
"y": 45
},
"id": 83,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "sum(rate(coredns_cache_hits_total[5m])) / (sum(rate(coredns_cache_hits_total[5m])) + sum(rate(coredns_cache_misses_total[5m]))) * 100",
"legendFormat": "",
"refId": "A"
}
],
"title": "DNS Cache Hit %",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"decimals": 1,
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "blue"
}
]
},
"unit": "reqps"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 12,
"y": 45
},
"id": 84,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "sum(rate(traefik_entrypoint_requests_total[5m]))",
"legendFormat": "",
"refId": "A"
}
],
"title": "Traefik req/s",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"decimals": 1,
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 5
},
{
"color": "red",
"value": 15
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 16,
"y": 45
},
"id": 85,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "sum(rate(traefik_entrypoint_requests_total{code=~\"[45]..\"}[5m])) / sum(rate(traefik_entrypoint_requests_total[5m])) * 100",
"legendFormat": "",
"refId": "A"
}
],
"title": "Traefik Error %",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"decimals": 1,
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "red"
},
{
"color": "orange",
"value": 50
},
{
"color": "green",
"value": 80
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 20,
"y": 45
},
"id": 86,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "sum(registry_registry_storage_cache_total{type=\"Hit\"}) / sum(registry_registry_storage_cache_total) * 100",
"legendFormat": "",
"refId": "A"
}
],
"title": "Registry Cache %",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"decimals": 1,
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "blue"
}
]
},
"unit": "reqps"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 8,
"y": 49
},
"id": 89,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "sum(rate(coredns_dns_requests_total[5m]))",
"legendFormat": "",
"refId": "A"
}
],
"title": "DNS req/s",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"decimals": 3,
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 0.1
},
{
"color": "red",
"value": 0.5
}
]
},
"unit": "s"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 12,
"y": 49
},
"id": 90,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "histogram_quantile(0.99, sum(rate(coredns_dns_request_duration_seconds_bucket[5m])) by (le))",
"legendFormat": "",
"refId": "A"
}
],
"title": "DNS Latency P99",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"decimals": 2,
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 1
},
{
"color": "red",
"value": 5
}
]
},
"unit": "s"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 16,
"y": 49
},
"id": 91,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "histogram_quantile(0.99, sum(rate(traefik_entrypoint_request_duration_seconds_bucket[5m])) by (le))",
"legendFormat": "",
"refId": "A"
}
],
"title": "Ingress Latency P99",
"type": "stat"
},
{
"collapsed": false,
"gridPos": {
"h": 1,
"w": 24,
"x": 0,
"y": 53
},
"id": 1,
"panels": [],
"title": "Cluster Summary",
"type": "row"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "red"
},
{
"color": "orange",
"value": 3
},
{
"color": "green",
"value": 5
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 0,
"y": 54
},
"id": 2,
"options": {
"colorMode": "value",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "sum(kube_node_status_condition{condition=\"Ready\",status=\"true\"})",
"legendFormat": "Ready",
"refId": "A"
}
],
"title": "Nodes Ready",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "blue"
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 4,
"y": 54
},
"id": 3,
"options": {
"colorMode": "value",
"graphMode": "area",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(kube_pod_info)",
"legendFormat": "Pods",
"refId": "A"
}
],
"title": "Total Pods",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [
{
"options": {
"0": {
"color": "green",
"text": "0"
}
},
"type": "value"
}
],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "red",
"value": 1
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 8,
"y": 54
},
"id": 4,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(ALERTS{alertstate=\"firing\"}) OR vector(0)",
"legendFormat": "Firing",
"refId": "A"
}
],
"title": "Firing Alerts",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 60
},
{
"color": "red",
"value": 80
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 12,
"y": 54
},
"id": 5,
"options": {
"minVizHeight": 75,
"minVizWidth": 75,
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showThresholdLabels": false,
"showThresholdMarkers": true
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "avg(1 - rate(node_cpu_seconds_total{mode=\"idle\"}[5m])) * 100",
"legendFormat": "CPU",
"refId": "A"
}
],
"title": "Cluster CPU %",
"type": "gauge"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 60
},
{
"color": "red",
"value": 80
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 16,
"y": 54
},
"id": 6,
"options": {
"minVizHeight": 75,
"minVizWidth": 75,
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showThresholdLabels": false,
"showThresholdMarkers": true
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "(1 - sum(node_memory_MemAvailable_bytes) / sum(node_memory_MemTotal_bytes)) * 100",
"legendFormat": "Memory",
"refId": "A"
}
],
"title": "Cluster Memory %",
"type": "gauge"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 70
},
{
"color": "red",
"value": 85
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 4,
"x": 20,
"y": 54
},
"id": 7,
"options": {
"minVizHeight": 75,
"minVizWidth": 75,
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showThresholdLabels": false,
"showThresholdMarkers": true
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "(1 - sum(node_filesystem_avail_bytes{mountpoint=\"/\",fstype!~\"tmpfs\"}) / sum(node_filesystem_size_bytes{mountpoint=\"/\",fstype!~\"tmpfs\"})) * 100",
"legendFormat": "Disk",
"refId": "A"
}
],
"title": "Cluster Disk %",
"type": "gauge"
},
{
"collapsed": false,
"gridPos": {
"h": 1,
"w": 24,
"x": 0,
"y": 58
},
"id": 20,
"panels": [],
"title": "Resource Governance — Quota Utilization",
"type": "row"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"custom": {
"align": "auto",
"cellOptions": {
"type": "auto"
},
"inspect": false
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 70
},
{
"color": "red",
"value": 90
}
]
}
},
"overrides": [
{
"matcher": {
"id": "byName",
"options": "Namespace"
},
"properties": [
{
"id": "custom.width",
"value": 160
}
]
},
{
"matcher": {
"id": "byName",
"options": "Resource"
},
"properties": [
{
"id": "custom.width",
"value": 120
}
]
},
{
"matcher": {
"id": "byName",
"options": "Usage %"
},
"properties": [
{
"id": "unit",
"value": "percent"
},
{
"id": "decimals",
"value": 1
},
{
"id": "custom.cellOptions",
"value": {
"mode": "gradient",
"type": "gauge"
}
}
]
}
]
},
"gridPos": {
"h": 8,
"w": 12,
"x": 0,
"y": 59
},
"id": 21,
"options": {
"cellHeight": "sm",
"footer": {
"countRows": false,
"fields": "",
"reducer": [
"sum"
],
"show": false
},
"showHeader": true,
"sortBy": [
{
"desc": true,
"displayName": "Usage %"
}
]
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "kube_resourcequota{type=\"used\"}",
"format": "table",
"instant": true,
"legendFormat": "",
"refId": "USED"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "kube_resourcequota{type=\"hard\"}",
"format": "table",
"instant": true,
"legendFormat": "",
"refId": "HARD"
}
],
"title": "Quota Usage by Namespace",
"transformations": [
{
"id": "merge",
"options": {}
},
{
"id": "calculateField",
"options": {
"alias": "Usage %",
"binary": {
"left": "Value #USED",
"operator": "/",
"reducer": "sum",
"right": "Value #HARD"
},
"mode": "binary",
"reduce": {
"reducer": "sum"
}
}
},
{
"id": "organize",
"options": {
"excludeByName": {
"Time": true,
"__name__": true,
"instance": true,
"job": true,
"resourcequota": true,
"type": true
},
"renameByName": {
"Usage %": "Usage %",
"Value #HARD": "Hard Limit",
"Value #USED": "Used",
"namespace": "Namespace",
"resource": "Resource"
}
}
},
{
"id": "filterByValue",
"options": {
"filters": [
{
"config": {
"id": "greater",
"options": {
"value": 0
}
},
"fieldName": "Hard Limit"
}
],
"match": "all",
"type": "include"
}
}
],
"type": "table"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 70
},
{
"color": "red",
"value": 90
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 8,
"w": 12,
"x": 12,
"y": 59
},
"id": 22,
"options": {
"displayMode": "gradient",
"maxVizHeight": 300,
"minVizHeight": 16,
"minVizWidth": 8,
"namePlacement": "auto",
"orientation": "horizontal",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showUnfilled": true,
"sizing": "auto",
"valueMode": "color"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "topk(15, kube_resourcequota{type=\"used\",resource=~\"requests.cpu|requests.memory|limits.cpu|limits.memory|pods\"} / kube_resourcequota{type=\"hard\",resource=~\"requests.cpu|requests.memory|limits.cpu|limits.memory|pods\"} * 100 > 0)",
"format": "table",
"instant": true,
"refId": "A"
}
],
"title": "Top Quota Consumers",
"transformations": [
{
"id": "organize",
"options": {
"excludeByName": {
"Time": true,
"__name__": true,
"endpoint": true,
"instance": true,
"job": true,
"metrics_path": true,
"resource": true,
"resourcequota": true,
"service": true,
"type": true,
"uid": true
},
"renameByName": {}
}
},
{
"id": "sortBy",
"options": {
"fields": {},
"sort": [
{
"desc": true,
"field": "Value"
}
]
}
},
{
"id": "rowsToFields",
"options": {}
}
],
"type": "bargauge"
},
{
"collapsed": false,
"gridPos": {
"h": 1,
"w": 24,
"x": 0,
"y": 67
},
"id": 30,
"panels": [],
"title": "Pod Health",
"type": "row"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "red",
"value": 1
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 0,
"y": 68
},
"id": 31,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(increase(kube_pod_container_status_restarts_total[1h]) > 5) OR vector(0)",
"legendFormat": "",
"refId": "A"
}
],
"title": "CrashLooping Pods",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 1
},
{
"color": "red",
"value": 5
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 3,
"y": 68
},
"id": 32,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "sum(increase(kube_pod_container_status_last_terminated_reason{reason=\"OOMKilled\"}[24h])) OR vector(0)",
"legendFormat": "",
"refId": "A"
}
],
"title": "OOMKilled (24h)",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "red",
"value": 1
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 6,
"y": 68
},
"id": 33,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(kube_pod_status_phase{phase=\"Failed\"}) OR vector(0)",
"legendFormat": "",
"refId": "A"
}
],
"title": "Failed Pods",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"noValue": "0",
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 1
},
{
"color": "red",
"value": 5
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 4,
"w": 3,
"x": 9,
"y": 68
},
"id": 34,
"options": {
"colorMode": "background",
"graphMode": "none",
"justifyMode": "auto",
"orientation": "auto",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"textMode": "auto"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "count(kube_pod_status_phase{phase=\"Pending\"}) OR vector(0)",
"legendFormat": "",
"refId": "A"
}
],
"title": "Pending Pods",
"type": "stat"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisBorderShow": false,
"axisCenteredZero": false,
"axisColorMode": "text",
"axisLabel": "",
"axisPlacement": "auto",
"barAlignment": 0,
"barWidthFactor": 0.6,
"drawStyle": "line",
"fillOpacity": 10,
"gradientMode": "none",
"hideFrom": {
"legend": false,
"tooltip": false,
"viz": false
},
"insertNulls": false,
"lineInterpolation": "smooth",
"lineWidth": 2,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"spanNulls": false,
"stacking": {
"group": "A",
"mode": "none"
},
"thresholdsStyle": {
"mode": "off"
}
},
"mappings": [],
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
}
]
}
},
"overrides": []
},
"gridPos": {
"h": 8,
"w": 12,
"x": 12,
"y": 68
},
"id": 35,
"options": {
"legend": {
"calcs": [
"sum"
],
"displayMode": "table",
"placement": "bottom",
"showLegend": true
},
"tooltip": {
"mode": "multi",
"sort": "desc"
}
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "sum(increase(kube_pod_container_status_restarts_total[1h]))",
"legendFormat": "Total Restarts / hour",
"refId": "A"
}
],
"title": "Restart Rate (24h)",
"type": "timeseries"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"custom": {
"align": "auto",
"cellOptions": {
"type": "auto"
},
"inspect": false
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 5
},
{
"color": "red",
"value": 20
}
]
}
},
"overrides": [
{
"matcher": {
"id": "byName",
"options": "Namespace"
},
"properties": [
{
"id": "custom.width",
"value": 140
}
]
},
{
"matcher": {
"id": "byName",
"options": "Pod"
},
"properties": [
{
"id": "custom.width",
"value": 260
}
]
},
{
"matcher": {
"id": "byName",
"options": "Restarts"
},
"properties": [
{
"id": "custom.cellOptions",
"value": {
"mode": "gradient",
"type": "gauge"
}
}
]
}
]
},
"gridPos": {
"h": 8,
"w": 12,
"x": 0,
"y": 72
},
"id": 36,
"options": {
"cellHeight": "sm",
"footer": {
"countRows": false,
"fields": "",
"reducer": [
"sum"
],
"show": false
},
"showHeader": true,
"sortBy": [
{
"desc": true,
"displayName": "Restarts"
}
]
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "topk(15, kube_pod_container_status_restarts_total)",
"format": "table",
"instant": true,
"legendFormat": "",
"refId": "A"
}
],
"title": "High-Restart Pods",
"transformations": [
{
"id": "organize",
"options": {
"excludeByName": {
"Time": true,
"__name__": true,
"container": true,
"endpoint": true,
"instance": true,
"job": true,
"service": true,
"uid": true
},
"renameByName": {
"Value": "Restarts",
"namespace": "Namespace",
"pod": "Pod"
}
}
}
],
"type": "table"
},
{
"collapsed": false,
"gridPos": {
"h": 1,
"w": 24,
"x": 0,
"y": 80
},
"id": 50,
"panels": [],
"title": "Storage",
"type": "row"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"custom": {
"align": "auto",
"cellOptions": {
"type": "auto"
},
"inspect": false
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 70
},
{
"color": "red",
"value": 85
}
]
}
},
"overrides": [
{
"matcher": {
"id": "byName",
"options": "Namespace"
},
"properties": [
{
"id": "custom.width",
"value": 140
}
]
},
{
"matcher": {
"id": "byName",
"options": "PVC"
},
"properties": [
{
"id": "custom.width",
"value": 260
}
]
},
{
"matcher": {
"id": "byName",
"options": "Used %"
},
"properties": [
{
"id": "unit",
"value": "percent"
},
{
"id": "decimals",
"value": 1
},
{
"id": "custom.cellOptions",
"value": {
"mode": "gradient",
"type": "gauge"
}
}
]
},
{
"matcher": {
"id": "byName",
"options": "Used"
},
"properties": [
{
"id": "unit",
"value": "bytes"
}
]
},
{
"matcher": {
"id": "byName",
"options": "Capacity"
},
"properties": [
{
"id": "unit",
"value": "bytes"
}
]
}
]
},
"gridPos": {
"h": 8,
"w": 12,
"x": 0,
"y": 81
},
"id": 51,
"options": {
"cellHeight": "sm",
"footer": {
"countRows": false,
"fields": "",
"reducer": [
"sum"
],
"show": false
},
"showHeader": true,
"sortBy": [
{
"desc": true,
"displayName": "Used %"
}
]
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "kubelet_volume_stats_used_bytes",
"format": "table",
"instant": true,
"legendFormat": "",
"refId": "USED"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "kubelet_volume_stats_capacity_bytes",
"format": "table",
"instant": true,
"legendFormat": "",
"refId": "CAP"
}
],
"title": "PV Usage Table",
"transformations": [
{
"id": "merge",
"options": {}
},
{
"id": "calculateField",
"options": {
"alias": "Used %",
"binary": {
"left": "Value #USED",
"operator": "/",
"right": "Value #CAP"
},
"mode": "binary"
}
},
{
"id": "organize",
"options": {
"excludeByName": {
"Time": true,
"__name__": true,
"endpoint": true,
"instance": true,
"job": true,
"metrics_path": true,
"node": true,
"service": true
},
"renameByName": {
"Used %": "Used %",
"Value #CAP": "Capacity",
"Value #USED": "Used",
"namespace": "Namespace",
"persistentvolumeclaim": "PVC"
}
}
},
{
"id": "filterByValue",
"options": {
"filters": [
{
"config": {
"id": "greater",
"options": {
"value": 0
}
},
"fieldName": "Capacity"
}
],
"match": "all",
"type": "include"
}
}
],
"type": "table"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "orange",
"value": 70
},
{
"color": "red",
"value": 85
}
]
},
"unit": "percent"
},
"overrides": []
},
"gridPos": {
"h": 8,
"w": 12,
"x": 12,
"y": 81
},
"id": 52,
"options": {
"displayMode": "gradient",
"maxVizHeight": 300,
"minVizHeight": 16,
"minVizWidth": 8,
"namePlacement": "auto",
"orientation": "horizontal",
"reduceOptions": {
"calcs": [
"lastNotNull"
],
"fields": "",
"values": false
},
"showUnfilled": true,
"sizing": "auto",
"valueMode": "color"
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "kubelet_volume_stats_used_bytes / kubelet_volume_stats_capacity_bytes * 100",
"legendFormat": "{{namespace}}/{{persistentvolumeclaim}}",
"refId": "A"
}
],
"title": "PV Usage Bar",
"type": "bargauge"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "palette-classic"
},
"custom": {
"axisBorderShow": false,
"axisCenteredZero": false,
"axisColorMode": "text",
"axisLabel": "",
"axisPlacement": "auto",
"barAlignment": 0,
"barWidthFactor": 0.6,
"drawStyle": "line",
"fillOpacity": 10,
"gradientMode": "none",
"hideFrom": {
"legend": false,
"tooltip": false,
"viz": false
},
"insertNulls": false,
"lineInterpolation": "smooth",
"lineWidth": 2,
"pointSize": 5,
"scaleDistribution": {
"type": "linear"
},
"showPoints": "never",
"spanNulls": true,
"stacking": {
"group": "A",
"mode": "none"
},
"thresholdsStyle": {
"mode": "off"
}
},
"mappings": [],
"max": 100,
"min": 0,
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "green"
},
{
"color": "red",
"value": 85
}
]
},
"unit": "percent"
},
"overrides": [
{
"matcher": {
"id": "byRegexp",
"options": "/^PVE /"
},
"properties": [
{
"id": "custom.lineStyle",
"value": {
"dash": [
10,
10
],
"fill": "dash"
}
},
{
"id": "custom.lineWidth",
"value": 2
}
]
}
]
},
"gridPos": {
"h": 8,
"w": 24,
"x": 0,
"y": 89
},
"id": 53,
"options": {
"legend": {
"calcs": [
"mean",
"max"
],
"displayMode": "table",
"placement": "bottom",
"showLegend": true
},
"tooltip": {
"mode": "multi",
"sort": "desc"
}
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "(1 - node_filesystem_avail_bytes{mountpoint=\"/\",fstype!~\"tmpfs\"} / node_filesystem_size_bytes{mountpoint=\"/\",fstype!~\"tmpfs\"}) * 100",
"legendFormat": "{{instance}}",
"refId": "A"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "pve_disk_usage_bytes{id=~\"storage/.*\"} / pve_disk_size_bytes{id=~\"storage/.*\"} * 100",
"legendFormat": "PVE {{id}}",
"refId": "B"
}
],
"title": "Node Disk Usage (24h)",
"type": "timeseries"
},
{
"collapsed": false,
"gridPos": {
"h": 1,
"w": 24,
"x": 0,
"y": 97
},
"id": 60,
"panels": [],
"title": "Certificate Expiry",
"type": "row"
},
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"fieldConfig": {
"defaults": {
"color": {
"mode": "thresholds"
},
"custom": {
"align": "auto",
"cellOptions": {
"type": "auto"
},
"inspect": false
},
"mappings": [],
"thresholds": {
"mode": "absolute",
"steps": [
{
"color": "red"
},
{
"color": "orange",
"value": 7
},
{
"color": "green",
"value": 30
}
]
}
},
"overrides": [
{
"matcher": {
"id": "byName",
"options": "Certificate"
},
"properties": [
{
"id": "custom.width",
"value": 400
}
]
},
{
"matcher": {
"id": "byName",
"options": "Days Remaining"
},
"properties": [
{
"id": "decimals",
"value": 0
},
{
"id": "custom.cellOptions",
"value": {
"mode": "gradient",
"type": "gauge"
}
}
]
}
]
},
"gridPos": {
"h": 8,
"w": 24,
"x": 0,
"y": 98
},
"id": 61,
"options": {
"cellHeight": "sm",
"footer": {
"countRows": false,
"fields": "",
"reducer": [
"sum"
],
"show": false
},
"showHeader": true,
"sortBy": [
{
"desc": false,
"displayName": "Days Remaining"
}
]
},
"targets": [
{
"datasource": {
"type": "prometheus",
"uid": "${datasource}"
},
"expr": "(traefik_tls_certs_not_after - time()) / 86400",
"format": "table",
"instant": true,
"legendFormat": "",
"refId": "A"
}
],
"title": "TLS Certificate Expiry",
"transformations": [
{
"id": "organize",
"options": {
"excludeByName": {
"Time": true,
"__name__": true,
"endpoint": true,
"instance": true,
"job": true,
"sans": true,
"service": true
},
"renameByName": {
"Value": "Days Remaining",
"cn": "Certificate",
"serial": "Serial"
}
}
},
{
"id": "sortBy",
"options": {
"fields": {},
"sort": [
{
"desc": false,
"field": "Days Remaining"
}
]
}
}
],
"type": "table"
}
],
"preload": false,
"refresh": "5m",
"schemaVersion": 42,
"tags": [
"cluster-health",
"daily-report"
],
"templating": {
"list": [
{
"current": {
"text": "Prometheus",
"value": "PBFA97CFB590B2093"
},
"includeAll": false,
"name": "datasource",
"options": [],
"query": "prometheus",
"refresh": 1,
"regex": "",
"type": "datasource"
}
]
},
"time": {
"from": "now-15m",
"to": "now"
},
"timepicker": {},
"timezone": "",
"title": "Cluster Health Overview",
"uid": "cluster-health-overview",
"version": 29
}