From 8838f2044f45451e2c7a86d374c2b3d39714405b Mon Sep 17 00:00:00 2001 From: Remigiusz Janeczek Date: Tue, 21 Apr 2020 15:40:52 +0200 Subject: Add script for loading prometheus snapshots, improve grafana 1. Add scripts for loading prometheus snapshots locally 2. Improve Grafana dashboards for performance tests 3. Improve Grafana dashboards management (now there's one configmap for all dashboards and to add new dashboard only needed change is adding new json file in grafana/dashboards directory) 4. Add prometheus job for scraping node metrics and add deployment yaml for node exporter Issue-ID: DCAEGEN2-1576 Signed-off-by: Remigiusz Janeczek Change-Id: I8870fd3291711b5dc09b76a7bf86ac4d13725237 --- .../cloud/grafana/dashboards/k8s-metrics.json | 530 +++++++++++++++++++++ 1 file changed, 530 insertions(+) create mode 100644 tools/performance/cloud/grafana/dashboards/k8s-metrics.json (limited to 'tools/performance/cloud/grafana/dashboards/k8s-metrics.json') diff --git a/tools/performance/cloud/grafana/dashboards/k8s-metrics.json b/tools/performance/cloud/grafana/dashboards/k8s-metrics.json new file mode 100644 index 00000000..676d0a93 --- /dev/null +++ b/tools/performance/cloud/grafana/dashboards/k8s-metrics.json @@ -0,0 +1,530 @@ + { + "annotations": { + "list": [ + { + "builtIn": 1, + "datasource": "-- Grafana --", + "enable": true, + "hide": true, + "iconColor": "rgba(0, 211, 255, 1)", + "name": "Annotations & Alerts", + "type": "dashboard" + } + ] + }, + "editable": true, + "gnetId": null, + "graphTooltip": 0, + "id": 5, + "links": [], + "panels": [ + { + "datasource": null, + "gridPos": { + "h": 7, + "w": 8, + "x": 0, + "y": 0 + }, + "id": 8, + "options": { + "fieldOptions": { + "calcs": [ + "mean" + ], + "defaults": { + "mappings": [], + "max": 100, + "min": 0, + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": null + }, + { + "color": "red", + "value": 80 + } + ] + }, + "unit": "percent" + }, + "overrides": [], + "values": false + }, + "orientation": "auto", + "showThresholdLabels": false, + "showThresholdMarkers": true + }, + "pluginVersion": "6.7.2", + "targets": [ + { + "expr": "sum(container_memory_working_set_bytes{id=\"/\", instance=~\".*worker.*\"}) by(instance) / (sum(machine_memory_bytes{instance=~\".*worker.*\"}) by (instance))*100", + "interval": "", + "legendFormat": "{{instance}}", + "refId": "A" + } + ], + "timeFrom": null, + "timeShift": null, + "title": "Nodes RAM Usage", + "type": "gauge" + }, + { + "datasource": null, + "gridPos": { + "h": 7, + "w": 8, + "x": 8, + "y": 0 + }, + "id": 6, + "options": { + "fieldOptions": { + "calcs": [ + "mean" + ], + "defaults": { + "mappings": [], + "max": 100, + "min": 0, + "thresholds": { + "mode": "absolute", + "steps": [ + { + "color": "green", + "value": null + }, + { + "color": "red", + "value": 80 + } + ] + }, + "unit": "percent" + }, + "overrides": [], + "values": false + }, + "orientation": "auto", + "showThresholdLabels": false, + "showThresholdMarkers": true + }, + "pluginVersion": "6.7.2", + "targets": [ + { + "expr": "sum(rate(container_cpu_usage_seconds_total{id=\"/\", instance=~\".*worker.*\"}[1m])) by(instance) / (sum(machine_cpu_cores{instance=~\".*worker.*\"}) by (instance))*100", + "interval": "", + "legendFormat": "{{instance}}", + "refId": "A" + } + ], + "timeFrom": null, + "timeShift": null, + "title": "Nodes CPU Usage", + "type": "gauge" + }, + { + "aliasColors": {}, + "bars": false, + "dashLength": 10, + "dashes": false, + "datasource": null, + "fill": 0, + "fillGradient": 0, + "gridPos": { + "h": 12, + "w": 8, + "x": 16, + "y": 0 + }, + "hiddenSeries": false, + "id": 10, + "legend": { + "avg": false, + "current": false, + "max": false, + "min": false, + "show": true, + "total": false, + "values": false + }, + "lines": true, + "linewidth": 1, + "nullPointMode": "null", + "options": { + "dataLinks": [] + }, + "percentage": false, + "pointradius": 2, + "points": false, + "renderer": "flot", + "seriesOverrides": [], + "spaceLength": 10, + "stack": false, + "steppedLine": false, + "targets": [ + { + "expr": "sum(irate(container_network_receive_bytes_total{pod!=\"\", namespace=\"onap\", instance=~\".*worker.*\"}[1m])) by (instance)", + "hide": false, + "interval": "", + "legendFormat": "{{instance}}", + "refId": "C" + }, + { + "expr": "sum(irate(container_network_receive_bytes_total{pod!=\"\", namespace=\"onap\", instance=~\".*worker.*\"}[1m]))", + "hide": false, + "interval": "", + "legendFormat": "Total Receive", + "refId": "D" + } + ], + "thresholds": [], + "timeFrom": "20m", + "timeRegions": [], + "timeShift": null, + "title": "Network Usage - Receive", + "tooltip": { + "shared": true, + "sort": 0, + "value_type": "individual" + }, + "type": "graph", + "xaxis": { + "buckets": null, + "mode": "time", + "name": null, + "show": true, + "values": [] + }, + "yaxes": [ + { + "format": "Bps", + "label": null, + "logBase": 1, + "max": null, + "min": null, + "show": true + }, + { + "format": "short", + "label": null, + "logBase": 1, + "max": null, + "min": null, + "show": true + } + ], + "yaxis": { + "align": false, + "alignLevel": null + } + }, + { + "aliasColors": {}, + "bars": false, + "dashLength": 10, + "dashes": false, + "datasource": null, + "fill": 0, + "fillGradient": 0, + "gridPos": { + "h": 16, + "w": 8, + "x": 0, + "y": 7 + }, + "hiddenSeries": false, + "id": 4, + "legend": { + "avg": false, + "current": false, + "max": false, + "min": false, + "show": true, + "total": false, + "values": false + }, + "lines": true, + "linewidth": 1, + "nullPointMode": "null", + "options": { + "dataLinks": [] + }, + "percentage": false, + "pointradius": 2, + "points": false, + "renderer": "flot", + "seriesOverrides": [], + "spaceLength": 10, + "stack": false, + "steppedLine": false, + "targets": [ + { + "expr": "sum(container_memory_working_set_bytes{id=\"/\"}) by (instance) / (sum(machine_memory_bytes) by (instance))*100", + "interval": "", + "legendFormat": "{{instance}}", + "refId": "A" + } + ], + "thresholds": [], + "timeFrom": "20m", + "timeRegions": [], + "timeShift": null, + "title": "Nodes Total RAM Usage", + "tooltip": { + "shared": true, + "sort": 0, + "value_type": "individual" + }, + "type": "graph", + "xaxis": { + "buckets": null, + "mode": "time", + "name": null, + "show": true, + "values": [] + }, + "yaxes": [ + { + "format": "percent", + "label": null, + "logBase": 1, + "max": null, + "min": null, + "show": true + }, + { + "format": "short", + "label": null, + "logBase": 1, + "max": null, + "min": null, + "show": true + } + ], + "yaxis": { + "align": false, + "alignLevel": null + } + }, + { + "aliasColors": {}, + "bars": false, + "dashLength": 10, + "dashes": false, + "datasource": null, + "fill": 0, + "fillGradient": 0, + "gridPos": { + "h": 16, + "w": 8, + "x": 8, + "y": 7 + }, + "hiddenSeries": false, + "id": 2, + "legend": { + "avg": false, + "current": false, + "max": false, + "min": false, + "show": true, + "total": false, + "values": false + }, + "lines": true, + "linewidth": 1, + "nullPointMode": "null", + "options": { + "dataLinks": [] + }, + "percentage": false, + "pointradius": 2, + "points": false, + "renderer": "flot", + "seriesOverrides": [], + "spaceLength": 10, + "stack": false, + "steppedLine": false, + "targets": [ + { + "expr": "sum(rate(container_cpu_usage_seconds_total{id=\"/\"}[1m])) by(instance) / (sum(machine_cpu_cores) by (instance))*100", + "interval": "", + "legendFormat": "{{instance}}", + "refId": "A" + } + ], + "thresholds": [], + "timeFrom": "20m", + "timeRegions": [], + "timeShift": null, + "title": "Nodes Total CPU usage", + "tooltip": { + "shared": true, + "sort": 0, + "value_type": "individual" + }, + "type": "graph", + "xaxis": { + "buckets": null, + "mode": "time", + "name": null, + "show": true, + "values": [] + }, + "yaxes": [ + { + "format": "percent", + "label": null, + "logBase": 1, + "max": null, + "min": null, + "show": true + }, + { + "format": "short", + "label": null, + "logBase": 1, + "max": null, + "min": null, + "show": true + } + ], + "yaxis": { + "align": false, + "alignLevel": null + } + }, + { + "aliasColors": {}, + "bars": false, + "dashLength": 10, + "dashes": false, + "datasource": null, + "fill": 0, + "fillGradient": 0, + "gridPos": { + "h": 11, + "w": 8, + "x": 16, + "y": 12 + }, + "hiddenSeries": false, + "id": 11, + "legend": { + "avg": false, + "current": false, + "max": false, + "min": false, + "show": true, + "total": false, + "values": false + }, + "lines": true, + "linewidth": 1, + "nullPointMode": "null", + "options": { + "dataLinks": [] + }, + "percentage": false, + "pointradius": 2, + "points": false, + "renderer": "flot", + "seriesOverrides": [], + "spaceLength": 10, + "stack": false, + "steppedLine": false, + "targets": [ + { + "expr": "sum(irate(container_network_transmit_bytes_total{pod!=\"\", namespace=\"onap\", instance=~\".*worker.*\"}[1m])) by (instance)", + "hide": false, + "interval": "", + "legendFormat": "{{instance}}", + "refId": "A" + }, + { + "expr": "sum(irate(container_network_transmit_bytes_total{pod!=\"\", namespace=\"onap\", instance=~\".*worker.*\"}[1m]))", + "hide": false, + "interval": "", + "legendFormat": "Total Transmit", + "refId": "B" + } + ], + "thresholds": [], + "timeFrom": "20m", + "timeRegions": [], + "timeShift": null, + "title": "Network Usage - Transmit", + "tooltip": { + "shared": true, + "sort": 0, + "value_type": "individual" + }, + "type": "graph", + "xaxis": { + "buckets": null, + "mode": "time", + "name": null, + "show": true, + "values": [] + }, + "yaxes": [ + { + "format": "Bps", + "label": null, + "logBase": 1, + "max": null, + "min": null, + "show": true + }, + { + "format": "short", + "label": null, + "logBase": 1, + "max": null, + "min": null, + "show": true + } + ], + "yaxis": { + "align": false, + "alignLevel": null + } + } + ], + "refresh": "5s", + "schemaVersion": 22, + "style": "dark", + "tags": [], + "templating": { + "list": [] + }, + "time": { + "from": "now-6h", + "to": "now" + }, + "timepicker": { + "refresh_intervals": [ + "5s", + "10s", + "30s", + "1m", + "5m", + "15m", + "30m", + "1h", + "2h", + "1d" + ] + }, + "timezone": "", + "title": "K8s metrics", + "uid": "RxjnWpjZk", + "variables": { + "list": [] + }, + "version": 1 + } -- cgit 1.2.3-korg