From 1a55967731935240077bc0a64f5316791e1a9207 Mon Sep 17 00:00:00 2001 From: brunomilitzer Date: Thu, 11 Aug 2022 16:16:57 +0100 Subject: Added Dashboards for Policy Framework Components Issue-ID: POLICY-3967 Change-Id: I33f135da7dc49dc10b42902d96d8da21db36d19b Signed-off-by: brunomilitzer --- .../metrics/dashboards/dashboard-distribution.json | 878 +++++++++++++++++++++ 1 file changed, 878 insertions(+) create mode 100644 csit/metrics/dashboards/dashboard-distribution.json (limited to 'csit/metrics/dashboards/dashboard-distribution.json') diff --git a/csit/metrics/dashboards/dashboard-distribution.json b/csit/metrics/dashboards/dashboard-distribution.json new file mode 100644 index 00000000..39b9675e --- /dev/null +++ b/csit/metrics/dashboards/dashboard-distribution.json @@ -0,0 +1,878 @@ +{ + "annotations": { + "list": [ + { + "builtIn": 1, + "datasource": "-- Grafana --", + "enable": true, + "hide": true, + "iconColor": "rgba(0, 211, 255, 1)", + "name": "Annotations & Alerts", + "target": { + "limit": 100, + "matchAny": false, + "tags": [], + "type": "dashboard" + }, + "type": "dashboard" + } + ] + }, + "description": "Grafana Dashboard for Policy Distribution", + "editable": true, + "fiscalYearStartMonth": 0, + "graphTooltip": 0, + "id": 8, + "links": [], + "liveNow": false, + "panels": [ + { + "collapsed": false, + "gridPos": { + "h": 1, + "w": 24, + "x": 0, + "y": 0 + }, + "id": 2, + "panels": [], + "title": "Quick Info", + "type": "row" + }, + { + "description": "Uptime per pod for Policy Distribution application calculated in days", + "fieldConfig": { + "defaults": { + "color": { + "mode": "thresholds" + }, + "mappings": [], + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": null + }, + { + "color": "red", + "value": 80 + } + ] + } + }, + "overrides": [] + }, + "gridPos": { + "h": 6, + "w": 6, + "x": 0, + "y": 1 + }, + "id": 4, + "options": { + "colorMode": "value", + "graphMode": "area", + "justifyMode": "auto", + "orientation": "auto", + "reduceOptions": { + "calcs": [], + "fields": "", + "values": false + }, + "textMode": "auto" + }, + "pluginVersion": "8.3.4", + "targets": [ + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "process_start_time_seconds{job=\"distribution-metrics\"} / 86400", + "interval": "", + "legendFormat": "", + "refId": "A" + } + ], + "title": "Process Uptime (days)", + "transparent": true, + "type": "stat" + }, + { + "description": "Policy Distribution CPU Usage Monitoring", + "fieldConfig": { + "defaults": { + "color": { + "mode": "thresholds" + }, + "mappings": [], + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": null + }, + { + "color": "red", + "value": 80 + } + ] + } + }, + "overrides": [] + }, + "gridPos": { + "h": 6, + "w": 8, + "x": 6, + "y": 1 + }, + "id": 6, + "options": { + "orientation": "auto", + "reduceOptions": { + "calcs": [ + "lastNotNull" + ], + "fields": "", + "values": false + }, + "showThresholdLabels": false, + "showThresholdMarkers": true + }, + "pluginVersion": "8.3.4", + "targets": [ + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "process_cpu_seconds_total{job=\"distribution-metrics\"}", + "interval": "", + "legendFormat": "process_cpu: {{pod}}", + "refId": "A" + } + ], + "title": "CPU Usage", + "type": "gauge" + }, + { + "description": "Policy Distribution Memory Usage Monitoring", + "fieldConfig": { + "defaults": { + "color": { + "mode": "thresholds" + }, + "mappings": [], + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": null + }, + { + "color": "red", + "value": 80 + } + ] + } + }, + "overrides": [] + }, + "gridPos": { + "h": 6, + "w": 8, + "x": 15, + "y": 1 + }, + "id": 8, + "options": { + "orientation": "auto", + "reduceOptions": { + "calcs": [ + "lastNotNull" + ], + "fields": "", + "values": false + }, + "showThresholdLabels": false, + "showThresholdMarkers": true + }, + "pluginVersion": "8.3.4", + "targets": [ + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "100 * sum by(pod)(jvm_memory_bytes_used{area=\"heap\", job=\"distribution-metrics\"}) / sum by(pod)(jvm_memory_bytes_max{area=\"heap\", job=\"distribution-metrics\"})", + "hide": false, + "interval": "", + "legendFormat": "Heap : {{pod}}", + "refId": "A" + }, + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "100 * sum by(pod)(jvm_memory_bytes_used{area=\"nonheap\", job=\"distribution-metrics\"}) / sum by(pod)(jvm_memory_bytes_max{area=\"nonheap\", job=\"distribution-metrics\"})", + "hide": false, + "interval": "", + "legendFormat": "Non Heap : {{pod}}", + "refId": "B" + } + ], + "title": "Memory Usage", + "type": "gauge" + }, + { + "collapsed": false, + "gridPos": { + "h": 1, + "w": 24, + "x": 0, + "y": 7 + }, + "id": 10, + "panels": [], + "title": "System", + "type": "row" + }, + { + "description": "Policy Distribution CPU Usage Monitoring", + "fieldConfig": { + "defaults": { + "color": { + "mode": "palette-classic" + }, + "custom": { + "axisLabel": "", + "axisPlacement": "auto", + "barAlignment": 0, + "drawStyle": "line", + "fillOpacity": 0, + "gradientMode": "none", + "hideFrom": { + "legend": false, + "tooltip": false, + "viz": false + }, + "lineInterpolation": "linear", + "lineWidth": 1, + "pointSize": 5, + "scaleDistribution": { + "type": "linear" + }, + "showPoints": "auto", + "spanNulls": false, + "stacking": { + "group": "A", + "mode": "none" + }, + "thresholdsStyle": { + "mode": "off" + } + }, + "mappings": [], + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": null + }, + { + "color": "red", + "value": 80 + } + ] + } + }, + "overrides": [] + }, + "gridPos": { + "h": 14, + "w": 12, + "x": 0, + "y": 8 + }, + "id": 12, + "options": { + "legend": { + "calcs": [], + "displayMode": "list", + "placement": "bottom" + }, + "tooltip": { + "mode": "single" + } + }, + "targets": [ + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "process_cpu_seconds_total{job=\"distribution-metrics\"} * 100", + "interval": "", + "legendFormat": "system: {{pod}}", + "refId": "A" + }, + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "process_cpu_seconds_total{job=\"distribution-metrics\"} * 100", + "hide": false, + "interval": "", + "legendFormat": "process: {{pod}}", + "refId": "B" + } + ], + "title": "CPU Usage", + "type": "timeseries" + }, + { + "description": "Policy Distribution Memory Usage Monitoring", + "fieldConfig": { + "defaults": { + "color": { + "mode": "palette-classic" + }, + "custom": { + "axisLabel": "", + "axisPlacement": "auto", + "barAlignment": 0, + "drawStyle": "line", + "fillOpacity": 0, + "gradientMode": "none", + "hideFrom": { + "legend": false, + "tooltip": false, + "viz": false + }, + "lineInterpolation": "linear", + "lineWidth": 1, + "pointSize": 5, + "scaleDistribution": { + "type": "linear" + }, + "showPoints": "auto", + "spanNulls": false, + "stacking": { + "group": "A", + "mode": "none" + }, + "thresholdsStyle": { + "mode": "off" + } + }, + "mappings": [], + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": null + }, + { + "color": "red", + "value": 80 + } + ] + } + }, + "overrides": [] + }, + "gridPos": { + "h": 8, + "w": 12, + "x": 12, + "y": 8 + }, + "id": 34, + "options": { + "legend": { + "calcs": [], + "displayMode": "list", + "placement": "bottom" + }, + "tooltip": { + "mode": "single" + } + }, + "targets": [ + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "100 * sum(jvm_memory_bytes_used{area=\"heap\", job=\"distribution-metrics\"}) / sum(jvm_memory_bytes_max{area=\"heap\", job=\"distribution-metrics\"})", + "interval": "", + "legendFormat": "Heap", + "refId": "A" + }, + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "100 * sum(jvm_memory_bytes_used{area=\"nonheap\", job=\"distribution-metrics\"}) / sum(jvm_memory_bytes_max{area=\"nonheap\", job=\"distribution-metrics\"})", + "hide": false, + "interval": "", + "legendFormat": "Non Heap", + "refId": "B" + } + ], + "title": "Memory Usage", + "type": "timeseries" + }, + { + "description": "Policy Distribution JVM Threads", + "fieldConfig": { + "defaults": { + "color": { + "mode": "palette-classic" + }, + "custom": { + "axisLabel": "", + "axisPlacement": "auto", + "barAlignment": 0, + "drawStyle": "line", + "fillOpacity": 0, + "gradientMode": "none", + "hideFrom": { + "legend": false, + "tooltip": false, + "viz": false + }, + "lineInterpolation": "linear", + "lineWidth": 1, + "pointSize": 5, + "scaleDistribution": { + "type": "linear" + }, + "showPoints": "auto", + "spanNulls": false, + "stacking": { + "group": "A", + "mode": "none" + }, + "thresholdsStyle": { + "mode": "off" + } + }, + "mappings": [], + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": null + }, + { + "color": "red", + "value": 80 + } + ] + } + }, + "overrides": [] + }, + "gridPos": { + "h": 13, + "w": 12, + "x": 12, + "y": 16 + }, + "id": 16, + "options": { + "legend": { + "calcs": [], + "displayMode": "list", + "placement": "bottom" + }, + "tooltip": { + "mode": "single" + } + }, + "targets": [ + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "jvm_threads_current{job=\"distribution-metrics\"}", + "interval": "", + "legendFormat": "Live Threads: {{ pod }}", + "refId": "A" + }, + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "jvm_threads_peak{job=\"distribution-metrics\"}", + "hide": false, + "interval": "", + "legendFormat": "Peak Threads: {{ pod }}", + "refId": "B" + } + ], + "title": "JVM threads", + "type": "timeseries" + }, + { + "collapsed": false, + "gridPos": { + "h": 1, + "w": 24, + "x": 0, + "y": 29 + }, + "id": 18, + "panels": [], + "title": "Garbage Collection", + "type": "row" + }, + { + "description": "GC operations per second", + "fieldConfig": { + "defaults": { + "color": { + "mode": "palette-classic" + }, + "custom": { + "axisLabel": "", + "axisPlacement": "auto", + "barAlignment": 0, + "drawStyle": "line", + "fillOpacity": 0, + "gradientMode": "none", + "hideFrom": { + "legend": false, + "tooltip": false, + "viz": false + }, + "lineInterpolation": "linear", + "lineWidth": 1, + "pointSize": 5, + "scaleDistribution": { + "type": "linear" + }, + "showPoints": "auto", + "spanNulls": false, + "stacking": { + "group": "A", + "mode": "none" + }, + "thresholdsStyle": { + "mode": "off" + } + }, + "mappings": [], + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": null + }, + { + "color": "red", + "value": 80 + } + ] + } + }, + "overrides": [] + }, + "gridPos": { + "h": 8, + "w": 12, + "x": 0, + "y": 30 + }, + "id": 20, + "options": { + "legend": { + "calcs": [], + "displayMode": "list", + "placement": "bottom" + }, + "tooltip": { + "mode": "single" + } + }, + "targets": [ + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "rate(jvm_gc_collection_seconds_count{job=\"distribution-metrics\"}[1h])", + "interval": "", + "legendFormat": "{{action}} ({{cause}}) : {{ pod }}", + "refId": "A" + } + ], + "title": "Collection", + "type": "timeseries" + }, + { + "description": "Average GC Time", + "fieldConfig": { + "defaults": { + "color": { + "mode": "palette-classic" + }, + "custom": { + "axisLabel": "", + "axisPlacement": "auto", + "barAlignment": 0, + "drawStyle": "line", + "fillOpacity": 0, + "gradientMode": "none", + "hideFrom": { + "legend": false, + "tooltip": false, + "viz": false + }, + "lineInterpolation": "linear", + "lineWidth": 1, + "pointSize": 5, + "scaleDistribution": { + "type": "linear" + }, + "showPoints": "auto", + "spanNulls": false, + "stacking": { + "group": "A", + "mode": "none" + }, + "thresholdsStyle": { + "mode": "off" + } + }, + "mappings": [], + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": null + }, + { + "color": "red", + "value": 80 + } + ] + } + }, + "overrides": [] + }, + "gridPos": { + "h": 8, + "w": 12, + "x": 12, + "y": 30 + }, + "id": 22, + "options": { + "legend": { + "calcs": [], + "displayMode": "list", + "placement": "bottom" + }, + "tooltip": { + "mode": "single" + } + }, + "targets": [ + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "rate(jvm_gc_collection_seconds_sum{job=\"distribution-metrics\"}[1h])/rate(jvm_gc_collection_seconds_count{job=\"distribution-metrics\"}[1h])", + "interval": "", + "legendFormat": "avg {{action}} ({{cause}}) : {{ pod }}", + "refId": "A" + } + ], + "title": "Pause Durations", + "type": "timeseries" + }, + { + "collapsed": false, + "gridPos": { + "h": 1, + "w": 24, + "x": 0, + "y": 38 + }, + "id": 24, + "panels": [], + "title": "Statistics", + "type": "row" + }, + { + "description": "Stats for Distribution Statistics Count", + "fieldConfig": { + "defaults": { + "color": { + "mode": "palette-classic" + }, + "mappings": [], + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": null + }, + { + "color": "red", + "value": 80 + } + ] + } + }, + "overrides": [] + }, + "gridPos": { + "h": 11, + "w": 12, + "x": 0, + "y": 39 + }, + "id": 32, + "options": { + "displayMode": "gradient", + "orientation": "vertical", + "reduceOptions": { + "calcs": [ + "lastNotNull" + ], + "fields": "", + "values": false + }, + "showUnfilled": false, + "text": { + "titleSize": 12 + } + }, + "pluginVersion": "8.3.4", + "targets": [ + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "sum(total_distribution_received_count{job=\"distribution-metrics\"})", + "interval": "", + "legendFormat": "Total Count", + "refId": "A" + }, + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "sum(distribution_success_count{job=\"distribution-metrics\"})", + "hide": false, + "interval": "", + "legendFormat": "Success Count", + "refId": "B" + }, + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "sum(distribution_failure_count{job=\"distribution-metrics\"})", + "hide": false, + "interval": "", + "legendFormat": "Failure Count", + "refId": "C" + }, + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "sum(total_download_received_count{job=\"distribution-metrics\"})", + "hide": false, + "interval": "", + "legendFormat": "Total Download Count", + "refId": "D" + }, + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "sum(download_success_count{job=\"distribution-metrics\"})", + "hide": false, + "interval": "", + "legendFormat": "Download Success Count", + "refId": "E" + }, + { + "datasource": { + "type": "prometheus", + "uid": "dkSf71fnz" + }, + "exemplar": true, + "expr": "sum(download_failure_count{job=\"distribution-metrics\"})", + "hide": false, + "interval": "", + "legendFormat": "Download Failure Count", + "refId": "F" + } + ], + "title": "Distribution Statistics", + "transparent": true, + "type": "bargauge" + } + ], + "refresh": "", + "schemaVersion": 34, + "style": "dark", + "tags": [], + "templating": { + "list": [] + }, + "time": { + "from": "now-6h", + "to": "now" + }, + "timepicker": {}, + "timezone": "", + "title": "Policy Distribution", + "uid": "5HkCopg4z", + "version": 2, + "weekStart": "" +} \ No newline at end of file -- cgit 1.2.3-korg