--- /dev/null
+{
+ "__requires": [
+ {
+ "type": "grafana",
+ "id": "grafana",
+ "name": "Grafana",
+ "version": "5.0.0"
+ },
+ {
+ "type": "panel",
+ "id": "graph",
+ "name": "Graph",
+ "version": "5.0.0"
+ },
+ {
+ "type": "panel",
+ "id": "singlestat",
+ "name": "Singlestat",
+ "version": "5.0.0"
+ }
+ ],
+ "annotations": {
+ "list": [
+ {
+ "builtIn": 1,
+ "datasource": "-- Grafana --",
+ "enable": true,
+ "hide": true,
+ "iconColor": "rgba(0, 211, 255, 1)",
+ "name": "Annotations & Alerts",
+ "type": "dashboard"
+ }
+ ]
+ },
+ "editable": false,
+ "gnetId": null,
+ "graphTooltip": 0,
+ "id": null,
+ "iteration": 1538079414024,
+ "links": [],
+ "panels": [
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "#299c46",
+ "rgba(237, 129, 40, 0.89)",
+ "#d44a3a"
+ ],
+ "datasource": null,
+ "format": "none",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 5,
+ "w": 4,
+ "x": 0,
+ "y": 0
+ },
+ "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,
+ "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(sum by (instance) (ceph_disk_occupation))",
+ "format": "time_series",
+ "instant": true,
+ "intervalFactor": 1,
+ "refId": "A"
+ }
+ ],
+ "thresholds": "",
+ "title": "OSD Hosts",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "#299c46",
+ "rgba(237, 129, 40, 0.89)",
+ "#d44a3a"
+ ],
+ "datasource": null,
+ "decimals": 0,
+ "description": "Average CPU busy across all hosts (OSD, RGW, MON etc) within the cluster",
+ "format": "percentunit",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 5,
+ "w": 4,
+ "x": 4,
+ "y": 0
+ },
+ "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,
+ "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": "avg(\n 1 - (\n avg by(instance) \n (irate(node_cpu_seconds_total{mode='idle',instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"}[1m]) or\n irate(node_cpu{mode='idle',instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"}[1m]))\n )\n )",
+ "format": "time_series",
+ "instant": true,
+ "intervalFactor": 1,
+ "refId": "A"
+ }
+ ],
+ "thresholds": "",
+ "title": "AVG CPU Busy",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "#299c46",
+ "rgba(237, 129, 40, 0.89)",
+ "#d44a3a"
+ ],
+ "datasource": null,
+ "decimals": 0,
+ "description": "Average Memory Usage across all hosts in the cluster (excludes buffer/cache usage)",
+ "format": "percentunit",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 5,
+ "w": 4,
+ "x": 8,
+ "y": 0
+ },
+ "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,
+ "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": "avg (((node_memory_MemTotal{instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"} or node_memory_MemTotal_bytes{instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"})- (\n (node_memory_MemFree{instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"} or node_memory_MemFree_bytes{instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"}) + \n (node_memory_Cached{instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"} or node_memory_Cached_bytes{instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"}) + \n (node_memory_Buffers{instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"} or node_memory_Buffers_bytes{instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"}) +\n (node_memory_Slab{instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"} or node_memory_Slab_bytes{instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"})\n )) /\n (node_memory_MemTotal{instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"} or node_memory_MemTotal_bytes{instance=~\"[[osd_hosts]]|[[rgw_hosts]]|[[mon_hosts]]|[[mds_hosts]].*\"} ))",
+ "format": "time_series",
+ "instant": true,
+ "intervalFactor": 1,
+ "refId": "A"
+ }
+ ],
+ "thresholds": "",
+ "title": "AVG RAM Utilization",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "#299c46",
+ "rgba(237, 129, 40, 0.89)",
+ "#d44a3a"
+ ],
+ "datasource": null,
+ "description": "IOPS Load at the device as reported by the OS on all OSD hosts",
+ "format": "none",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 5,
+ "w": 4,
+ "x": 12,
+ "y": 0
+ },
+ "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,
+ "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 ((irate(node_disk_reads_completed{instance=~\"($osd_hosts).*\"}[5m]) or irate(node_disk_reads_completed_total{instance=~\"($osd_hosts).*\"}[5m]) ) + \n(irate(node_disk_writes_completed{instance=~\"($osd_hosts).*\"}[5m]) or irate(node_disk_writes_completed_total{instance=~\"($osd_hosts).*\"}[5m])))",
+ "format": "time_series",
+ "instant": true,
+ "intervalFactor": 1,
+ "refId": "A"
+ }
+ ],
+ "thresholds": "",
+ "title": "Physical IOPS",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "#299c46",
+ "rgba(237, 129, 40, 0.89)",
+ "#d44a3a"
+ ],
+ "datasource": null,
+ "description": "Average Disk utilization for all OSD data devices (i.e. excludes journal/WAL)",
+ "format": "percent",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 5,
+ "w": 4,
+ "x": 16,
+ "y": 0
+ },
+ "id": 20,
+ "interval": null,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "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": "avg (\n ((irate(node_disk_io_time_ms[5m]) / 10 ) or\n (irate(node_disk_io_time_seconds_total[5m]) * 100)\n ) *\n on(instance, device) ceph_disk_occupation{instance=~\"($osd_hosts).*\"}\n)",
+ "format": "time_series",
+ "instant": true,
+ "intervalFactor": 1,
+ "refId": "A"
+ }
+ ],
+ "thresholds": "",
+ "title": "AVG Disk Utilization",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "cacheTimeout": null,
+ "colorBackground": false,
+ "colorValue": false,
+ "colors": [
+ "#299c46",
+ "rgba(237, 129, 40, 0.89)",
+ "#d44a3a"
+ ],
+ "datasource": null,
+ "decimals": 0,
+ "description": "Total send/receive network load across all hosts in the ceph cluster",
+ "format": "bytes",
+ "gauge": {
+ "maxValue": 100,
+ "minValue": 0,
+ "show": false,
+ "thresholdLabels": false,
+ "thresholdMarkers": true
+ },
+ "gridPos": {
+ "h": 5,
+ "w": 4,
+ "x": 20,
+ "y": 0
+ },
+ "id": 18,
+ "interval": null,
+ "links": [],
+ "mappingType": 1,
+ "mappingTypes": [
+ {
+ "name": "value to text",
+ "value": 1
+ },
+ {
+ "name": "range to text",
+ "value": 2
+ }
+ ],
+ "maxDataPoints": 100,
+ "nullPointMode": "connected",
+ "nullText": null,
+ "postfix": "",
+ "postfixFontSize": "50%",
+ "prefix": "",
+ "prefixFontSize": "50%",
+ "rangeMaps": [
+ {
+ "from": "null",
+ "text": "N/A",
+ "to": "null"
+ }
+ ],
+ "sparkline": {
+ "fillColor": "rgba(31, 118, 189, 0.18)",
+ "full": false,
+ "lineColor": "rgb(31, 120, 193)",
+ "show": false
+ },
+ "tableColumn": "",
+ "targets": [
+ {
+ "expr": "sum (\n irate(node_network_receive_bytes{instance=~\"($osd_hosts|mon_hosts|mds_hosts|rgw_hosts).*\",device!=\"lo\"}[1m]) or\n irate(node_network_receive_bytes_total{instance=~\"($osd_hosts|mon_hosts|mds_hosts|rgw_hosts).*\",device!=\"lo\"}[1m])\n ) +\nsum (\n irate(node_network_transmit_bytes{instance=~\"($osd_hosts|mon_hosts|mds_hosts|rgw_hosts).*\",device!=\"lo\"}[1m]) or\n irate(node_network_transmit_bytes_total{instance=~\"($osd_hosts|mon_hosts|mds_hosts|rgw_hosts).*\",device!=\"lo\"}[1m])\n )",
+ "format": "time_series",
+ "instant": true,
+ "intervalFactor": 1,
+ "refId": "A"
+ }
+ ],
+ "thresholds": "",
+ "title": "Network Load",
+ "type": "singlestat",
+ "valueFontSize": "80%",
+ "valueMaps": [
+ {
+ "op": "=",
+ "text": "N/A",
+ "value": "null"
+ }
+ ],
+ "valueName": "current"
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "description": "Show the top 10 busiest hosts by cpu",
+ "fill": 1,
+ "gridPos": {
+ "h": 9,
+ "w": 12,
+ "x": 0,
+ "y": 5
+ },
+ "id": 13,
+ "legend": {
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "show": false,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 1,
+ "links": [],
+ "nullPointMode": "null",
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": true,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "topk(10,( 1 - (\n avg by(instance) \n (irate(node_cpu_seconds_total{mode='idle',instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"}[1m]) or\n irate(node_cpu{mode='idle',instance=~\"($osd_hosts|$mon_hosts|$mds_hosts|$rgw_hosts).*\"}[1m]))\n )\n )\n)",
+ "format": "time_series",
+ "intervalFactor": 1,
+ "legendFormat": "{{instance}}",
+ "refId": "A"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "CPU Busy - Top 10 Hosts",
+ "tooltip": {
+ "shared": true,
+ "sort": 1,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "decimals": 1,
+ "format": "percentunit",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": "0",
+ "show": false
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ]
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "description": "Top 10 hosts by network load",
+ "fill": 1,
+ "gridPos": {
+ "h": 9,
+ "w": 12,
+ "x": 12,
+ "y": 5
+ },
+ "id": 19,
+ "legend": {
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "show": false,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 1,
+ "links": [],
+ "nullPointMode": "null",
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": true,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "topk(10, (sum by(instance) (\n (\n irate(node_network_receive_bytes{instance=~\"($osd_hosts|mon_hosts|mds_hosts|rgw_hosts).*\",device!=\"lo\"}[1m]) or\n irate(node_network_receive_bytes_total{instance=~\"($osd_hosts|mon_hosts|mds_hosts|rgw_hosts).*\",device!=\"lo\"}[1m])\n ) +\n (\n irate(node_network_transmit_bytes{instance=~\"($osd_hosts|mon_hosts|mds_hosts|rgw_hosts).*\",device!=\"lo\"}[1m]) or\n irate(node_network_transmit_bytes_total{instance=~\"($osd_hosts|mon_hosts|mds_hosts|rgw_hosts).*\",device!=\"lo\"}[1m])\n ))\n )\n)",
+ "format": "time_series",
+ "intervalFactor": 1,
+ "legendFormat": "{{instance}}",
+ "refId": "A"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "Network Load - Top 10",
+ "tooltip": {
+ "shared": true,
+ "sort": 1,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "decimals": 1,
+ "format": "bytes",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": "0",
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ]
+ }
+ ],
+ "refresh": "10s",
+ "schemaVersion": 16,
+ "style": "dark",
+ "tags": [],
+ "templating": {
+ "list": [
+ {
+ "allValue": "",
+ "current": {},
+ "datasource": null,
+ "hide": 2,
+ "includeAll": true,
+ "label": null,
+ "multi": false,
+ "name": "osd_hosts",
+ "options": [],
+ "query": "label_values(ceph_disk_occupation, instance)",
+ "refresh": 1,
+ "regex": "([^.]*).*",
+ "sort": 1,
+ "tagValuesQuery": "",
+ "tags": [],
+ "tagsQuery": "ceph",
+ "type": "query",
+ "useTags": false
+ },
+ {
+ "allValue": null,
+ "current": {},
+ "datasource": null,
+ "hide": 2,
+ "includeAll": true,
+ "label": null,
+ "multi": false,
+ "name": "mon_hosts",
+ "options": [],
+ "query": "label_values(ceph_mon_metadata, ceph_daemon)",
+ "refresh": 1,
+ "regex": "mon.(.*)",
+ "sort": 0,
+ "tagValuesQuery": "",
+ "tags": [],
+ "tagsQuery": "",
+ "type": "query",
+ "useTags": false
+ },
+ {
+ "allValue": null,
+ "current": {},
+ "datasource": null,
+ "hide": 2,
+ "includeAll": true,
+ "label": null,
+ "multi": false,
+ "name": "mds_hosts",
+ "options": [],
+ "query": "label_values(ceph_mds_inodes, ceph_daemon)",
+ "refresh": 1,
+ "regex": "mds.(.*)",
+ "sort": 0,
+ "tagValuesQuery": "",
+ "tags": [],
+ "tagsQuery": "",
+ "type": "query",
+ "useTags": false
+ },
+ {
+ "allValue": null,
+ "current": {},
+ "datasource": null,
+ "hide": 2,
+ "includeAll": true,
+ "label": null,
+ "multi": false,
+ "name": "rgw_hosts",
+ "options": [],
+ "query": "label_values(ceph_rgw_qlen, ceph_daemon)",
+ "refresh": 1,
+ "regex": "rgw.(.*)",
+ "sort": 0,
+ "tagValuesQuery": "",
+ "tags": [],
+ "tagsQuery": "",
+ "type": "query",
+ "useTags": false
+ }
+ ]
+ },
+ "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": "",
+ "title": "Host Overview",
+ "uid": "lxnjcTAmk",
+ "version": 10
+}
--- /dev/null
+{
+
+ "__requires": [
+ {
+ "type": "grafana",
+ "id": "grafana",
+ "name": "Grafana",
+ "version": "5.0.0"
+ },
+ {
+ "type": "panel",
+ "id": "grafana-piechart-panel",
+ "name": "Pie Chart",
+ "version": "1.3.3"
+ },
+ {
+ "type": "panel",
+ "id": "graph",
+ "name": "Graph",
+ "version": "5.0.0"
+ },
+ {
+ "type": "panel",
+ "id": "table",
+ "name": "Table",
+ "version": "5.0.0"
+ }
+ ],
+ "annotations": {
+ "list": [
+ {
+ "builtIn": 1,
+ "datasource": "-- Grafana --",
+ "enable": true,
+ "hide": true,
+ "iconColor": "rgba(0, 211, 255, 1)",
+ "name": "Annotations & Alerts",
+ "type": "dashboard"
+ }
+ ]
+ },
+ "editable": false,
+ "gnetId": null,
+ "graphTooltip": 0,
+ "id": null,
+ "iteration": 1538083987689,
+ "links": [],
+ "panels": [
+ {
+ "aliasColors": {
+ "@95%ile": "#e0752d"
+ },
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "fill": 1,
+ "gridPos": {
+ "h": 8,
+ "w": 8,
+ "x": 0,
+ "y": 0
+ },
+ "id": 12,
+ "legend": {
+ "avg": false,
+ "current": true,
+ "max": false,
+ "min": false,
+ "show": true,
+ "total": false,
+ "values": true
+ },
+ "lines": true,
+ "linewidth": 1,
+ "links": [],
+ "nullPointMode": "null",
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "avg (irate(ceph_osd_op_r_latency_sum[1m]) / on (ceph_daemon) irate(ceph_osd_op_r_latency_count[1m]) * 1000)",
+ "format": "time_series",
+ "intervalFactor": 1,
+ "legendFormat": "AVG read",
+ "refId": "A"
+ },
+ {
+ "expr": "max (irate(ceph_osd_op_r_latency_sum[1m]) / on (ceph_daemon) irate(ceph_osd_op_r_latency_count[1m]) * 1000)",
+ "format": "time_series",
+ "intervalFactor": 1,
+ "legendFormat": "MAX read",
+ "refId": "B"
+ },
+ {
+ "expr": "quantile(0.95,\n (irate(ceph_osd_op_r_latency_sum[1m]) / on (ceph_daemon) irate(ceph_osd_op_r_latency_count[1m]) * 1000)\n)",
+ "format": "time_series",
+ "intervalFactor": 1,
+ "legendFormat": "@95%ile",
+ "refId": "C"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "OSD Read Latencies",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "format": "ms",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": "0",
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ]
+ },
+ {
+ "columns": [],
+ "datasource": null,
+ "description": "This table shows the osd's that are delivering the 10 highest read latencies within the cluster",
+ "fontSize": "100%",
+ "gridPos": {
+ "h": 8,
+ "w": 4,
+ "x": 8,
+ "y": 0
+ },
+ "id": 15,
+ "links": [],
+ "pageSize": null,
+ "scroll": true,
+ "showHeader": true,
+ "sort": {
+ "col": 2,
+ "desc": true
+ },
+ "styles": [
+ {
+ "alias": "OSD ID",
+ "colorMode": null,
+ "colors": [
+ "rgba(245, 54, 54, 0.9)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(50, 172, 45, 0.97)"
+ ],
+ "dateFormat": "YYYY-MM-DD HH:mm:ss",
+ "decimals": 2,
+ "pattern": "ceph_daemon",
+ "thresholds": [],
+ "type": "string",
+ "unit": "short"
+ },
+ {
+ "alias": "Latency (ms)",
+ "colorMode": null,
+ "colors": [
+ "rgba(245, 54, 54, 0.9)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(50, 172, 45, 0.97)"
+ ],
+ "dateFormat": "YYYY-MM-DD HH:mm:ss",
+ "decimals": 0,
+ "pattern": "Value",
+ "thresholds": [],
+ "type": "number",
+ "unit": "none"
+ },
+ {
+ "alias": "",
+ "colorMode": null,
+ "colors": [
+ "rgba(245, 54, 54, 0.9)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(50, 172, 45, 0.97)"
+ ],
+ "dateFormat": "YYYY-MM-DD HH:mm:ss",
+ "decimals": 2,
+ "pattern": "/.*/",
+ "thresholds": [],
+ "type": "hidden",
+ "unit": "short"
+ }
+ ],
+ "targets": [
+ {
+ "expr": "topk(10,\n (sort(\n (irate(ceph_osd_op_r_latency_sum[1m]) / on (ceph_daemon) irate(ceph_osd_op_r_latency_count[1m]) * 1000)\n ))\n)\n\n",
+ "format": "table",
+ "instant": true,
+ "intervalFactor": 1,
+ "legendFormat": "",
+ "refId": "A"
+ }
+ ],
+ "title": "Highest READ Latencies",
+ "transform": "table",
+ "type": "table"
+ },
+ {
+ "aliasColors": {
+ "@95%ile write": "#e0752d"
+ },
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "fill": 1,
+ "gridPos": {
+ "h": 8,
+ "w": 8,
+ "x": 12,
+ "y": 0
+ },
+ "id": 13,
+ "legend": {
+ "avg": false,
+ "current": true,
+ "max": false,
+ "min": false,
+ "show": true,
+ "total": false,
+ "values": true
+ },
+ "lines": true,
+ "linewidth": 1,
+ "links": [],
+ "nullPointMode": "null",
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "avg (irate(ceph_osd_op_w_latency_sum[1m]) / on (ceph_daemon) irate(ceph_osd_op_w_latency_count[1m]) * 1000)",
+ "format": "time_series",
+ "intervalFactor": 1,
+ "legendFormat": "AVG write",
+ "refId": "A"
+ },
+ {
+ "expr": "max (irate(ceph_osd_op_w_latency_sum[1m]) / on (ceph_daemon) irate(ceph_osd_op_w_latency_count[1m]) * 1000)",
+ "format": "time_series",
+ "intervalFactor": 1,
+ "legendFormat": "MAX write",
+ "refId": "B"
+ },
+ {
+ "expr": "quantile(0.95,\n (irate(ceph_osd_op_w_latency_sum[1m]) / on (ceph_daemon) irate(ceph_osd_op_w_latency_count[1m]) * 1000)\n)",
+ "format": "time_series",
+ "intervalFactor": 1,
+ "legendFormat": "@95%ile write",
+ "refId": "C"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "OSD Write Latencies",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": []
+ },
+ "yaxes": [
+ {
+ "format": "ms",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": "0",
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ]
+ },
+ {
+ "columns": [],
+ "datasource": null,
+ "description": "This table shows the osd's that are delivering the 10 highest write latencies within the cluster",
+ "fontSize": "100%",
+ "gridPos": {
+ "h": 8,
+ "w": 4,
+ "x": 20,
+ "y": 0
+ },
+ "id": 16,
+ "links": [],
+ "pageSize": null,
+ "scroll": true,
+ "showHeader": true,
+ "sort": {
+ "col": 2,
+ "desc": true
+ },
+ "styles": [
+ {
+ "alias": "OSD ID",
+ "colorMode": null,
+ "colors": [
+ "rgba(245, 54, 54, 0.9)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(50, 172, 45, 0.97)"
+ ],
+ "dateFormat": "YYYY-MM-DD HH:mm:ss",
+ "decimals": 2,
+ "pattern": "ceph_daemon",
+ "thresholds": [],
+ "type": "string",
+ "unit": "short"
+ },
+ {
+ "alias": "Latency (ms)",
+ "colorMode": null,
+ "colors": [
+ "rgba(245, 54, 54, 0.9)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(50, 172, 45, 0.97)"
+ ],
+ "dateFormat": "YYYY-MM-DD HH:mm:ss",
+ "decimals": 0,
+ "pattern": "Value",
+ "thresholds": [],
+ "type": "number",
+ "unit": "none"
+ },
+ {
+ "alias": "",
+ "colorMode": null,
+ "colors": [
+ "rgba(245, 54, 54, 0.9)",
+ "rgba(237, 129, 40, 0.89)",
+ "rgba(50, 172, 45, 0.97)"
+ ],
+ "dateFormat": "YYYY-MM-DD HH:mm:ss",
+ "decimals": 2,
+ "pattern": "/.*/",
+ "thresholds": [],
+ "type": "hidden",
+ "unit": "short"
+ }
+ ],
+ "targets": [
+ {
+ "expr": "topk(10,\n (sort(\n (irate(ceph_osd_op_w_latency_sum[1m]) / on (ceph_daemon) irate(ceph_osd_op_w_latency_count[1m]) * 1000)\n ))\n)\n\n",
+ "format": "table",
+ "instant": true,
+ "intervalFactor": 1,
+ "legendFormat": "",
+ "refId": "A"
+ }
+ ],
+ "title": "Highest WRITE Latencies",
+ "transform": "table",
+ "type": "table"
+ },
+ {
+ "aliasColors": {},
+ "breakPoint": "50%",
+ "cacheTimeout": null,
+ "combine": {
+ "label": "Others",
+ "threshold": 0
+ },
+ "datasource": null,
+ "fontSize": "80%",
+ "format": "none",
+ "gridPos": {
+ "h": 8,
+ "w": 4,
+ "x": 0,
+ "y": 8
+ },
+ "id": 2,
+ "interval": null,
+ "legend": {
+ "show": true,
+ "values": true
+ },
+ "legendType": "Under graph",
+ "links": [],
+ "maxDataPoints": 3,
+ "nullPointMode": "connected",
+ "pieType": "pie",
+ "strokeWidth": 1,
+ "targets": [
+ {
+ "expr": "count by(device_class) (ceph_osd_metadata)",
+ "format": "time_series",
+ "intervalFactor": 1,
+ "legendFormat": "{{device_class}}",
+ "refId": "A"
+ }
+ ],
+ "title": "OSD Types Summary",
+ "type": "grafana-piechart-panel",
+ "valueName": "current"
+ },
+ {
+ "aliasColors": {
+ "Non-Encrypted": "#E5AC0E"
+ },
+ "breakPoint": "50%",
+ "cacheTimeout": null,
+ "combine": {
+ "label": "Others",
+ "threshold": 0
+ },
+ "datasource": null,
+ "fontSize": "80%",
+ "format": "none",
+ "gridPos": {
+ "h": 8,
+ "w": 4,
+ "x": 4,
+ "y": 8
+ },
+ "height": "200px",
+ "hideTimeOverride": true,
+ "id": 4,
+ "interval": null,
+ "legend": {
+ "percentage": false,
+ "show": true,
+ "values": true
+ },
+ "legendType": "Under graph",
+ "links": [],
+ "maxDataPoints": "1",
+ "minSpan": 4,
+ "nullPointMode": "connected",
+ "pieType": "pie",
+ "strokeWidth": 1,
+ "targets": [
+ {
+ "expr": "count(ceph_bluefs_wal_total_bytes)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "bluestore",
+ "refId": "A",
+ "step": 240
+ },
+ {
+ "expr": "count(ceph_osd_metadata) - count(ceph_bluefs_wal_total_bytes)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "filestore",
+ "refId": "B",
+ "step": 240
+ },
+ {
+ "expr": "absent(ceph_bluefs_wal_total_bytes)*count(ceph_osd_metadata)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "filestore",
+ "refId": "C",
+ "step": 240
+ }
+ ],
+ "timeFrom": "2m",
+ "timeShift": null,
+ "title": "OSD Objectstore Types",
+ "type": "grafana-piechart-panel",
+ "valueName": "current"
+ },
+ {
+ "aliasColors": {},
+ "breakPoint": "50%",
+ "cacheTimeout": null,
+ "combine": {
+ "label": "Others",
+ "threshold": "0.05"
+ },
+ "datasource": null,
+ "description": "The pie chart shows the various OSD sizes used within the cluster",
+ "fontSize": "80%",
+ "format": "none",
+ "gridPos": {
+ "h": 8,
+ "w": 4,
+ "x": 8,
+ "y": 8
+ },
+ "height": "220",
+ "hideTimeOverride": true,
+ "id": 8,
+ "interval": null,
+ "legend": {
+ "header": "",
+ "percentage": false,
+ "show": true,
+ "sideWidth": null,
+ "sortDesc": true,
+ "values": true
+ },
+ "legendType": "Under graph",
+ "links": [],
+ "maxDataPoints": "",
+ "minSpan": 6,
+ "nullPointMode": "connected",
+ "pieType": "pie",
+ "strokeWidth": "1",
+ "targets": [
+ {
+ "expr": "count(ceph_osd_stat_bytes < 1099511627776)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "<1 TB",
+ "refId": "A",
+ "step": 2
+ },
+ {
+ "expr": "count(ceph_osd_stat_bytes >= 1099511627776 < 2199023255552)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "<2 TB",
+ "refId": "B",
+ "step": 2
+ },
+ {
+ "expr": "count(ceph_osd_stat_bytes >= 2199023255552 < 3298534883328)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "<3TB",
+ "refId": "C",
+ "step": 2
+ },
+ {
+ "expr": "count(ceph_osd_stat_bytes >= 3298534883328 < 4398046511104)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "<4TB",
+ "refId": "D",
+ "step": 2
+ },
+ {
+ "expr": "count(ceph_osd_stat_bytes >= 4398046511104 < 6597069766656)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "<6TB",
+ "refId": "E",
+ "step": 2
+ },
+ {
+ "expr": "count(ceph_osd_stat_bytes >= 6597069766656 < 8796093022208)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "<8TB",
+ "refId": "F",
+ "step": 2
+ },
+ {
+ "expr": "count(ceph_osd_stat_bytes >= 8796093022208 < 10995116277760)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "<10TB",
+ "refId": "G",
+ "step": 2
+ },
+ {
+ "expr": "count(ceph_osd_stat_bytes >= 10995116277760 < 13194139533312)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "<12TB",
+ "refId": "H",
+ "step": 2
+ },
+ {
+ "expr": "count(ceph_osd_stat_bytes >= 13194139533312)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "12TB+",
+ "refId": "I",
+ "step": 2
+ }
+ ],
+ "timeFrom": "2m",
+ "timeShift": null,
+ "title": "OSD Size Summary",
+ "type": "grafana-piechart-panel",
+ "valueName": "current"
+ },
+ {
+ "aliasColors": {},
+ "bars": true,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "description": "Each bar indicates the number of OSD's that have a PG count in a specific range as shown on the x axis.",
+ "fill": 1,
+ "gridPos": {
+ "h": 8,
+ "w": 12,
+ "x": 12,
+ "y": 8
+ },
+ "id": 6,
+ "legend": {
+ "alignAsTable": false,
+ "avg": false,
+ "current": false,
+ "hideEmpty": false,
+ "hideZero": false,
+ "max": false,
+ "min": false,
+ "rightSide": false,
+ "show": false,
+ "total": false,
+ "values": false
+ },
+ "lines": false,
+ "linewidth": 1,
+ "links": [],
+ "nullPointMode": "null",
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "ceph_osd_numpg\n",
+ "format": "time_series",
+ "instant": true,
+ "intervalFactor": 1,
+ "legendFormat": "PGs per OSD",
+ "refId": "A"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "Distribution of PGs per OSD",
+ "tooltip": {
+ "shared": false,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": 20,
+ "mode": "histogram",
+ "name": null,
+ "show": true,
+ "values": [
+ "total"
+ ]
+ },
+ "yaxes": [
+ {
+ "decimals": 0,
+ "format": "short",
+ "label": "# of OSDs",
+ "logBase": 1,
+ "max": null,
+ "min": "0",
+ "show": true
+ },
+ {
+ "format": "short",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": false
+ }
+ ]
+ },
+ {
+ "collapsed": false,
+ "gridPos": {
+ "h": 1,
+ "w": 24,
+ "x": 0,
+ "y": 16
+ },
+ "id": 20,
+ "panels": [],
+ "title": "R/W Profile",
+ "type": "row"
+ },
+ {
+ "aliasColors": {},
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": null,
+ "description": "Show the read/write workload profile overtime",
+ "fill": 1,
+ "gridPos": {
+ "h": 8,
+ "w": 24,
+ "x": 0,
+ "y": 17
+ },
+ "id": 10,
+ "legend": {
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "show": true,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 1,
+ "links": [],
+ "nullPointMode": "null",
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "seriesOverrides": [],
+ "spaceLength": 10,
+ "stack": true,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "round(sum(irate(ceph_pool_rd[30s])))",
+ "format": "time_series",
+ "intervalFactor": 1,
+ "legendFormat": "Reads",
+ "refId": "A"
+ },
+ {
+ "expr": "round(sum(irate(ceph_pool_wr[30s])))",
+ "format": "time_series",
+ "intervalFactor": 1,
+ "legendFormat": "Writes",
+ "refId": "B"
+ }
+ ],
+ "thresholds": [],
+ "timeFrom": "36h",
+ "timeShift": null,
+ "title": "Read/Write Profile",
+ "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
+ }
+ ]
+ }
+ ],
+ "refresh": "10s",
+ "schemaVersion": 16,
+ "style": "dark",
+ "tags": [],
+ "templating": {
+ "list": []
+ },
+ "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": "",
+ "title": "OSD Overview",
+ "uid": "lo02I1Aiz",
+ "version": 3
+}
projects / ceph-ci.git / commitdiff