[unifi-poller] metrics refactor (#1182)

Co-authored-by: ᗪєνιη ᗷυнʟ <onedr0p@users.noreply.github.com>

charts/stable/unifi-poller/Chart.yaml

@@ -2,7 +2,7 @@ apiVersion: v2
 appVersion: 2.1.3
 description: Collect ALL UniFi Controller, Site, Device & Client Data - Export to InfluxDB or Prometheus
 name: unifi-poller
-version: 9.0.0
+version: 10.0.0
 kubeVersion: ">=1.16.0-0"
 keywords:
 - unifi

charts/stable/unifi-poller/README.md

@@ -1,6 +1,6 @@
 # unifi-poller
 
-![Version: 9.0.0](https://img.shields.io/badge/Version-9.0.0-informational?style=flat-square) ![AppVersion: 2.1.3](https://img.shields.io/badge/AppVersion-2.1.3-informational?style=flat-square)
+![Version: 10.0.0](https://img.shields.io/badge/Version-10.0.0-informational?style=flat-square) ![AppVersion: 2.1.3](https://img.shields.io/badge/AppVersion-2.1.3-informational?style=flat-square)
 
 Collect ALL UniFi Controller, Site, Device & Client Data - Export to InfluxDB or Prometheus
 
@@ -87,10 +87,16 @@ N/A
 | influxdb.enabled | bool | `false` | Create an InfluxDB instance as a [unifi-poller storage backend](https://unifipoller.com/docs/dependencies/influxdb).    See [bitnami/influxdb](https://github.com/bitnami/charts/tree/master/bitnami/influxdb) for more options. |
 | influxdb.persistence.enabled | bool | `false` | Enable persistence to store in a PV so data survives pod restarts. |
 | ingress.main.enabled | bool | `false` | Expose [unifi-poller's web interface](https://unifipoller.com/docs/advanced/webserver)    (if enabled in the configuration) via the k8s ingress by setting this true. |
-| prometheus.serviceMonitor.additionalLabels | object | `{}` | More labels to add to the ServiceMonitor resource. |
+| metrics.enabled | bool | See values.yaml | Enable and configure a Prometheus serviceMonitor for the chart under this key. |
-| prometheus.serviceMonitor.enabled | bool | `false` | Create a [ServiceMonitor](https://github.com/prometheus-operator/prometheus-operator/blob/master/Documentation/design.md#servicemonitor)    for the [unifi-poller prometheus exporter](https://unifipoller.com/docs/dependencies/prometheus)    that is recognized by [prometheus-operator](https://github.com/prometheus-operator/prometheus-operator).    This gets created in this chart's namespace. Note: this only creates a ServiceMonitor, not a prometheus instance. |
+| metrics.prometheusRule | object | See values.yaml | Enable and configure Prometheus Rules for the chart under this key. |
-| prometheus.serviceMonitor.interval | string | `"1m"` | Prometheus [endpoint polling interval](https://github.com/prometheus-operator/prometheus-operator/blob/master/Documentation/api.md#endpoint)    for this ServiceMonitor (how often unifi-poller will be queried). |
+| metrics.prometheusRule.rules | list | See prometheusrules.yaml | Configure additionial rules for the chart under this key. |
-| service.main.ports.http.port | int | `9130` | The port this prometheus exporter will listen on ([registered as 9130](https://github.com/prometheus/prometheus/wiki/Default-port-allocations)).    This should match unifi-poller's [configured listen port](https://unifipoller.com/docs/install/configuration#prometheus). |
+| metrics.serviceMonitor.interval | string | `"30s"` |  |
+| metrics.serviceMonitor.labels | object | `{}` |  |
+| metrics.serviceMonitor.scrapeTimeout | string | `"10s"` |  |
+| service.main.ports.http.enabled | bool | `false` |  |
+| service.main.ports.metrics.enabled | bool | `true` |  |
+| service.main.ports.metrics.port | int | `9130` |  |
+| service.main.ports.metrics.protocol | string | `"TCP"` |  |
 
 ## Changelog
 
@@ -98,6 +104,12 @@ All notable changes to this application Helm chart will be documented in this fi
 
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+### [10.0.0]
+
+#### Changed
+
+- **BREAKING**: Refactored Prometheus metrics section to add rules. Enabling metrics automatically enables the serviceMonitor.
+
 ### [9.0.0]
 
 #### Changed
@@ -113,6 +125,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - **BREAKING**: Change the default influxdb database name to match unifi-poller's default
 - Updates default unifi-poller version from 2.0.1 to 2.1.3
 
+[10.0.0]: #1000
 [9.0.0]: #900
 [8.0.0]: #800
 

charts/stable/unifi-poller/README_CHANGELOG.md.gotmpl

@@ -9,6 +9,12 @@ All notable changes to this application Helm chart will be documented in this fi
 
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+### [10.0.0]
+
+#### Changed
+
+- **BREAKING**: Refactored Prometheus metrics section to add rules. Enabling metrics automatically enables the serviceMonitor.
+
 ### [9.0.0]
 
 #### Changed
@@ -24,6 +30,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - **BREAKING**: Change the default influxdb database name to match unifi-poller's default
 - Updates default unifi-poller version from 2.0.1 to 2.1.3
 
+[10.0.0]: #1000
 [9.0.0]: #900
 [8.0.0]: #800
 {{- end -}}

charts/stable/unifi-poller/templates/prometheusrules.yaml

@@ -0,0 +1,27 @@
+{{- if and .Values.metrics.enabled .Values.metrics.prometheusRule.enabled }}
+apiVersion: monitoring.coreos.com/v1
+kind: PrometheusRule
+metadata:
+  name: {{ include "common.names.fullname" . }}
+  labels:
+    {{- include "common.labels" . | nindent 4 }}
+    {{- with .Values.metrics.prometheusRule.labels }}
+    {{- toYaml . | nindent 4 }}
+    {{- end }}
+spec:
+  groups:
+    - name: {{ include "common.names.fullname" . }}
+      rules:
+        - alert: UnifiPollerAbsent
+          annotations:
+            description: Unifi Poller has disappeared from Prometheus service discovery.
+            summary: Unifi Poller is down.
+          expr: |
+            absent(up{job=~".*{{ include "common.names.fullname" . }}.*"} == 1)
+          for: 5m
+          labels:
+            severity: critical
+        {{- with .Values.metrics.prometheusRule.rules }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+{{- end }}

charts/stable/unifi-poller/templates/servicemonitor.yaml

@@ -1,11 +1,11 @@
-{{- if .Values.prometheus.serviceMonitor.enabled }}
+{{- if .Values.metrics.enabled }}
 apiVersion: monitoring.coreos.com/v1
 kind: ServiceMonitor
 metadata:
-  name: {{ template "common.names.fullname" . }}
+  name: {{ include "common.names.fullname" . }}
   labels:
     {{- include "common.labels" . | nindent 4 }}
-    {{- with .Values.prometheus.serviceMonitor.additionalLabels }}
+    {{- with .Values.metrics.serviceMonitor.labels }}
     {{- toYaml . | nindent 4 }}
     {{- end }}
 spec:
@@ -13,9 +13,12 @@ spec:
     matchLabels:
       {{- include "common.labels.selectorLabels" . | nindent 6 }}
   endpoints:
-  - port: http
+    - port: metrics
-    {{- with .Values.prometheus.serviceMonitor.interval }}
+      {{- with .Values.metrics.serviceMonitor.interval }}
-    interval: {{ . }}
+      interval: {{ . }}
-    {{- end }}
+      {{- end }}
-    path: /metrics
+      {{- with .Values.metrics.serviceMonitor.scrapeTimeout }}
+      scrapeTimeout: {{ . }}
+      {{- end }}
+      path: /metrics
 {{- end }}

charts/stable/unifi-poller/values.yaml

@@ -28,8 +28,10 @@ service:
   main:
     ports:
       http:
-        # -- The port this prometheus exporter will listen on ([registered as 9130](https://github.com/prometheus/prometheus/wiki/Default-port-allocations)).
+        enabled: false
-        #    This should match unifi-poller's [configured listen port](https://unifipoller.com/docs/install/configuration#prometheus).
+      metrics:
+        enabled: true
+        protocol: TCP
         port: 9130
 
 ingress:
@@ -38,19 +40,6 @@ ingress:
     #    (if enabled in the configuration) via the k8s ingress by setting this true.
     enabled: false
 
-prometheus:
-  serviceMonitor:
-    # -- Create a [ServiceMonitor](https://github.com/prometheus-operator/prometheus-operator/blob/master/Documentation/design.md#servicemonitor)
-    #    for the [unifi-poller prometheus exporter](https://unifipoller.com/docs/dependencies/prometheus)
-    #    that is recognized by [prometheus-operator](https://github.com/prometheus-operator/prometheus-operator).
-    #    This gets created in this chart's namespace. Note: this only creates a ServiceMonitor, not a prometheus instance.
-    enabled: false
-    # -- Prometheus [endpoint polling interval](https://github.com/prometheus-operator/prometheus-operator/blob/master/Documentation/api.md#endpoint)
-    #    for this ServiceMonitor (how often unifi-poller will be queried).
-    interval: 1m
-    # -- More labels to add to the ServiceMonitor resource.
-    additionalLabels: {}
-
 influxdb:
   # -- Create an InfluxDB instance as a [unifi-poller storage backend](https://unifipoller.com/docs/dependencies/influxdb).
   #    See [bitnami/influxdb](https://github.com/bitnami/charts/tree/master/bitnami/influxdb) for more options.
@@ -68,3 +57,29 @@ influxdb:
     enabled: false
     # storageClass: ""
     # size: 8Gi
+
+metrics:
+  # -- Enable and configure a Prometheus serviceMonitor for the chart under this key.
+  # @default -- See values.yaml
+  enabled: false
+  serviceMonitor:
+    interval: 1m
+    scrapeTimeout: 30s
+    labels: {}
+  # -- Enable and configure Prometheus Rules for the chart under this key.
+  # @default -- See values.yaml
+  prometheusRule:
+    enabled: false
+    labels: {}
+    # -- Configure additionial rules for the chart under this key.
+    # @default -- See prometheusrules.yaml
+    rules: []
+      # - alert: UnifiPollerAbsent
+      #   annotations:
+      #     description: Unifi Poller has disappeared from Prometheus service discovery.
+      #     summary: Unifi Poller is down.
+      #   expr: |
+      #     absent(up{job=~".*unifi-poller.*"} == 1)
+      #   for: 5m
+      #   labels:
+      #     severity: critical