diff --git a/jsonnet/kube-prometheus/main.libsonnet b/jsonnet/kube-prometheus/main.libsonnet
index e1dc0b5955c2e065eaac9abe98b14852088316df..0873e7e89526fad11cee96c67a8a9184a0af7cca 100644
--- a/jsonnet/kube-prometheus/main.libsonnet
+++ b/jsonnet/kube-prometheus/main.libsonnet
@@ -68,7 +68,7 @@ local platformPatch = import './platforms/platforms.libsonnet';
image: $.values.common.images.grafana,
prometheusName: $.values.prometheus.name,
// TODO(paulfantom) This should be done by iterating over all objects and looking for object.mixin.grafanaDashboards
- dashboards: $.nodeExporter.mixin.grafanaDashboards + $.prometheus.mixin.grafanaDashboards + $.kubernetesControlPlane.mixin.grafanaDashboards,
+ dashboards: $.nodeExporter.mixin.grafanaDashboards + $.prometheus.mixin.grafanaDashboards + $.kubernetesControlPlane.mixin.grafanaDashboards + $.alertmanager.mixin.grafanaDashboards,
},
kubeStateMetrics: {
namespace: $.values.common.namespace,
diff --git a/jsonnetfile.lock.json b/jsonnetfile.lock.json
index a7b5053d47b8bf4e2add4eb674c78e825b3f5748..f338b75264e56ec81c2b61a0e407d2f639271a4c 100644
--- a/jsonnetfile.lock.json
+++ b/jsonnetfile.lock.json
@@ -109,8 +109,8 @@
"subdir": "doc/alertmanager-mixin"
}
},
- "version": "7301451eb94d2081fb740abf29755861dd122c65",
- "sum": "V8jcZQ1Qrlm7AQ6wjbuQQsacPb0NvrcZovKyplmzW5w=",
+ "version": "8598683b2461fb68e1921735c20163c4c784f9b6",
+ "sum": "YIWuR6x64SRQSCr8tuuGN1cc0TK5HGR0HWvgot3fc6k=",
"name": "alertmanager"
},
{
diff --git a/manifests/grafana-dashboardDefinitions.yaml b/manifests/grafana-dashboardDefinitions.yaml
index 267433dbc4f05e9c1c30078c87de4001e1d117e6..de6813ea4f7887770db4c042ee9a19341fa11363 100644
--- a/manifests/grafana-dashboardDefinitions.yaml
+++ b/manifests/grafana-dashboardDefinitions.yaml
@@ -1,5 +1,604 @@
apiVersion: v1
items:
+- apiVersion: v1
+ data:
+ alertmanager-overview.json: |-
+ {
+ "__inputs": [
+
+ ],
+ "__requires": [
+
+ ],
+ "annotations": {
+ "list": [
+
+ ]
+ },
+ "editable": false,
+ "gnetId": null,
+ "graphTooltip": 1,
+ "hideControls": false,
+ "id": null,
+ "links": [
+
+ ],
+ "refresh": "30s",
+ "rows": [
+ {
+ "collapse": false,
+ "collapsed": false,
+ "panels": [
+ {
+ "aliasColors": {
+
+ },
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": "$datasource",
+ "fill": 1,
+ "fillGradient": 0,
+ "gridPos": {
+
+ },
+ "id": 2,
+ "legend": {
+ "alignAsTable": false,
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "rightSide": false,
+ "show": false,
+ "sideWidth": null,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 1,
+ "links": [
+
+ ],
+ "nullPointMode": "null",
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "repeat": null,
+ "seriesOverrides": [
+
+ ],
+ "spaceLength": 10,
+ "span": 6,
+ "stack": true,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "sum(alertmanager_alerts{namespace=\"$namespace\",service=\"$service\"}) by (namespace,service,instance)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "{{instance}}",
+ "refId": "A"
+ }
+ ],
+ "thresholds": [
+
+ ],
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "Alerts",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": [
+
+ ]
+ },
+ "yaxes": [
+ {
+ "format": "none",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "none",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ }
+ ]
+ },
+ {
+ "aliasColors": {
+
+ },
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": "$datasource",
+ "fill": 1,
+ "fillGradient": 0,
+ "gridPos": {
+
+ },
+ "id": 3,
+ "legend": {
+ "alignAsTable": false,
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "rightSide": false,
+ "show": false,
+ "sideWidth": null,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 1,
+ "links": [
+
+ ],
+ "nullPointMode": "null",
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "repeat": null,
+ "seriesOverrides": [
+
+ ],
+ "spaceLength": 10,
+ "span": 6,
+ "stack": true,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "sum(rate(alertmanager_alerts_received_total{namespace=\"$namespace\",service=\"$service\"}[5m])) by (namespace,service,instance)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "{{instance}} Received",
+ "refId": "A"
+ },
+ {
+ "expr": "sum(rate(alertmanager_alerts_invalid_total{namespace=\"$namespace\",service=\"$service\"}[5m])) by (namespace,service,instance)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "{{instance}} Invalid",
+ "refId": "B"
+ }
+ ],
+ "thresholds": [
+
+ ],
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "Alerts receive rate",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": [
+
+ ]
+ },
+ "yaxes": [
+ {
+ "format": "ops",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "ops",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ }
+ ]
+ }
+ ],
+ "repeat": null,
+ "repeatIteration": null,
+ "repeatRowId": null,
+ "showTitle": true,
+ "title": "Alerts",
+ "titleSize": "h6",
+ "type": "row"
+ },
+ {
+ "collapse": false,
+ "collapsed": false,
+ "panels": [
+ {
+ "aliasColors": {
+
+ },
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": "$datasource",
+ "fill": 1,
+ "fillGradient": 0,
+ "gridPos": {
+
+ },
+ "id": 4,
+ "legend": {
+ "alignAsTable": false,
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "rightSide": false,
+ "show": false,
+ "sideWidth": null,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 1,
+ "links": [
+
+ ],
+ "nullPointMode": "null",
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "repeat": "integration",
+ "seriesOverrides": [
+
+ ],
+ "spaceLength": 10,
+ "stack": true,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "sum(rate(alertmanager_notifications_total{namespace=\"$namespace\",service=\"$service\", integration=\"$integration\"}[5m])) by (integration,namespace,service,instance)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "{{instance}} Total",
+ "refId": "A"
+ },
+ {
+ "expr": "sum(rate(alertmanager_notifications_failed_total{namespace=\"$namespace\",service=\"$service\", integration=\"$integration\"}[5m])) by (integration,namespace,service,instance)",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "{{instance}} Failed",
+ "refId": "B"
+ }
+ ],
+ "thresholds": [
+
+ ],
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "$integration: Notifications Send Rate",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": [
+
+ ]
+ },
+ "yaxes": [
+ {
+ "format": "ops",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "ops",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ }
+ ]
+ },
+ {
+ "aliasColors": {
+
+ },
+ "bars": false,
+ "dashLength": 10,
+ "dashes": false,
+ "datasource": "$datasource",
+ "fill": 1,
+ "fillGradient": 0,
+ "gridPos": {
+
+ },
+ "id": 5,
+ "legend": {
+ "alignAsTable": false,
+ "avg": false,
+ "current": false,
+ "max": false,
+ "min": false,
+ "rightSide": false,
+ "show": false,
+ "sideWidth": null,
+ "total": false,
+ "values": false
+ },
+ "lines": true,
+ "linewidth": 1,
+ "links": [
+
+ ],
+ "nullPointMode": "null",
+ "percentage": false,
+ "pointradius": 5,
+ "points": false,
+ "renderer": "flot",
+ "repeat": "integration",
+ "seriesOverrides": [
+
+ ],
+ "spaceLength": 10,
+ "stack": false,
+ "steppedLine": false,
+ "targets": [
+ {
+ "expr": "histogram_quantile(0.99,\n sum(rate(alertmanager_notification_latency_seconds_bucket{namespace=\"$namespace\",service=\"$service\", integration=\"$integration\"}[5m])) by (le,namespace,service,instance)\n) \n",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "{{instance}} 99th Percentile",
+ "refId": "A"
+ },
+ {
+ "expr": "histogram_quantile(0.50,\n sum(rate(alertmanager_notification_latency_seconds_bucket{namespace=\"$namespace\",service=\"$service\", integration=\"$integration\"}[5m])) by (le,namespace,service,instance)\n) \n",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "{{instance}} Median",
+ "refId": "B"
+ },
+ {
+ "expr": "sum(rate(alertmanager_notification_latency_seconds_sum{namespace=\"$namespace\",service=\"$service\", integration=\"$integration\"}[5m])) by (namespace,service,instance)\n/\nsum(rate(alertmanager_notification_latency_seconds_count{namespace=\"$namespace\",service=\"$service\", integration=\"$integration\"}[5m])) by (namespace,service,instance)\n",
+ "format": "time_series",
+ "intervalFactor": 2,
+ "legendFormat": "{{instance}} Average",
+ "refId": "C"
+ }
+ ],
+ "thresholds": [
+
+ ],
+ "timeFrom": null,
+ "timeShift": null,
+ "title": "$integration: Notification Duration",
+ "tooltip": {
+ "shared": true,
+ "sort": 0,
+ "value_type": "individual"
+ },
+ "type": "graph",
+ "xaxis": {
+ "buckets": null,
+ "mode": "time",
+ "name": null,
+ "show": true,
+ "values": [
+
+ ]
+ },
+ "yaxes": [
+ {
+ "format": "s",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ },
+ {
+ "format": "s",
+ "label": null,
+ "logBase": 1,
+ "max": null,
+ "min": null,
+ "show": true
+ }
+ ]
+ }
+ ],
+ "repeat": null,
+ "repeatIteration": null,
+ "repeatRowId": null,
+ "showTitle": true,
+ "title": "Notifications",
+ "titleSize": "h6",
+ "type": "row"
+ }
+ ],
+ "schemaVersion": 14,
+ "style": "dark",
+ "tags": [
+ "alertmanager-mixin"
+ ],
+ "templating": {
+ "list": [
+ {
+ "current": {
+ "text": "Prometheus",
+ "value": "Prometheus"
+ },
+ "hide": 0,
+ "label": null,
+ "name": "datasource",
+ "options": [
+
+ ],
+ "query": "prometheus",
+ "refresh": 1,
+ "regex": "",
+ "type": "datasource"
+ },
+ {
+ "allValue": null,
+ "current": {
+ "text": "",
+ "value": ""
+ },
+ "datasource": "$datasource",
+ "hide": 0,
+ "includeAll": false,
+ "label": null,
+ "multi": false,
+ "name": "namespace",
+ "options": [
+
+ ],
+ "query": "label_values(alertmanager_alerts, namespace)",
+ "refresh": 2,
+ "regex": "",
+ "sort": 1,
+ "tagValuesQuery": "",
+ "tags": [
+
+ ],
+ "tagsQuery": "",
+ "type": "query",
+ "useTags": false
+ },
+ {
+ "allValue": null,
+ "current": {
+ "text": "",
+ "value": ""
+ },
+ "datasource": "$datasource",
+ "hide": 0,
+ "includeAll": false,
+ "label": null,
+ "multi": false,
+ "name": "service",
+ "options": [
+
+ ],
+ "query": "label_values(alertmanager_alerts, service)",
+ "refresh": 2,
+ "regex": "",
+ "sort": 1,
+ "tagValuesQuery": "",
+ "tags": [
+
+ ],
+ "tagsQuery": "",
+ "type": "query",
+ "useTags": false
+ },
+ {
+ "allValue": null,
+ "current": {
+ "text": "all",
+ "value": "$__all"
+ },
+ "datasource": "$datasource",
+ "hide": 2,
+ "includeAll": true,
+ "label": null,
+ "multi": false,
+ "name": "integration",
+ "options": [
+
+ ],
+ "query": "label_values(alertmanager_notifications_total{integration=~\".*\"}, integration)",
+ "refresh": 2,
+ "regex": "",
+ "sort": 1,
+ "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": "utc",
+ "title": "Alertmanager / Overview",
+ "uid": "alertmanager-overview",
+ "version": 0
+ }
+ kind: ConfigMap
+ metadata:
+ labels:
+ app.kubernetes.io/component: grafana
+ app.kubernetes.io/name: grafana
+ app.kubernetes.io/part-of: kube-prometheus
+ app.kubernetes.io/version: 7.5.4
+ name: grafana-dashboard-alertmanager-overview
+ namespace: monitoring
- apiVersion: v1
data:
apiserver.json: |-
diff --git a/manifests/grafana-deployment.yaml b/manifests/grafana-deployment.yaml
index edb6623ea45f0c681f5075f68e79d5d6e819f6e5..c889337bf3e32aad8000e3e0a1aad1a7b7574b30 100644
--- a/manifests/grafana-deployment.yaml
+++ b/manifests/grafana-deployment.yaml
@@ -53,6 +53,9 @@ spec:
- mountPath: /etc/grafana/provisioning/dashboards
name: grafana-dashboards
readOnly: false
+ - mountPath: /grafana-dashboard-definitions/0/alertmanager-overview
+ name: grafana-dashboard-alertmanager-overview
+ readOnly: false
- mountPath: /grafana-dashboard-definitions/0/apiserver
name: grafana-dashboard-apiserver
readOnly: false
@@ -135,6 +138,9 @@ spec:
- configMap:
name: grafana-dashboards
name: grafana-dashboards
+ - configMap:
+ name: grafana-dashboard-alertmanager-overview
+ name: grafana-dashboard-alertmanager-overview
- configMap:
name: grafana-dashboard-apiserver
name: grafana-dashboard-apiserver