feat(helm): ability to monitor garage via prometheus

script/helm/garage/templates/service.yaml

@@ -17,3 +17,24 @@ spec:
       name: s3-web
   selector:
     {{- include "garage.selectorLabels" . | nindent 4 }}
+{{- if .Values.monitoring.metrics.enabled }}
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: {{ include "garage.fullname" . }}-metrics
+  labels:
+    {{- include "garage.labels" . | nindent 4 }}
+  annotations:
+    prometheus.io/scrape: "true"
+spec:
+  type: ClusterIP
+  clusterIP: None
+  ports:
+    - port: 3903
+      targetPort: 3903
+      protocol: TCP
+      name: metrics
+  selector:
+    {{- include "garage.selectorLabels" . | nindent 4 }}
+{{- end }}

script/helm/garage/templates/servicemonitor.yaml

@@ -0,0 +1,44 @@
+{{- if .Values.monitoring.metrics.serviceMonitor.enabled }}
+---
+apiVersion: monitoring.coreos.com/v1
+kind: ServiceMonitor
+metadata:
+  name: {{ include "garage.fullname" . }}
+  {{- if .Values.monitoring.metrics.serviceMonitor.namespace }}
+  namespace: {{ tpl .Values.monitoring.metrics.serviceMonitor.namespace . }}
+  {{- else }}
+  namespace: {{ .Release.Namespace }}
+  {{- end }}
+  labels:
+    {{- include "garage.labels" . | nindent 4 }}
+    {{- with .Values.monitoring.metrics.serviceMonitor.labels }}
+    {{- toYaml . | nindent 4 }}
+    {{- end }}
+spec:
+  endpoints:
+    - port: metrics
+    {{- with .Values.monitoring.metrics.serviceMonitor.interval }}
+      interval: {{ . }}
+    {{- end }}
+    {{- with .Values.monitoring.metrics.serviceMonitor.scrapeTimeout }}
+      scrapeTimeout: {{ . }}
+    {{- end }}
+      honorLabels: true
+      path: {{ .Values.monitoring.metrics.serviceMonitor.path }}
+      scheme: {{ .Values.monitoring.metrics.serviceMonitor.scheme }}
+    {{- with .Values.monitoring.metrics.serviceMonitor.tlsConfig }}
+      tlsConfig:
+      {{- toYaml . | nindent 6 }}
+    {{- end }}
+    {{- with .Values.monitoring.metrics.serviceMonitor.relabelings }}
+      relabelings:
+      {{- toYaml . | nindent 6 }}
+    {{- end }}
+  jobLabel: "{{ .Release.Name }}"
+  selector:
+    matchLabels:
+      {{- include "garage.selectorLabels" . | nindent 6 }}
+  namespaceSelector:
+    matchNames:
+      - {{ .Release.Namespace }}
+{{- end }}

script/helm/garage/templates/workload.yaml

@@ -61,6 +61,8 @@ spec:
               name: s3-api
             - containerPort: 3902
               name: web-api
+            - containerPort: 3903
+              name: admin
           volumeMounts:
             - name: meta
               mountPath: /mnt/meta

script/helm/garage/values.yaml

@@ -49,6 +49,12 @@ garage:
     bind_addr = "[::]:3902"
     root_domain = "{{ .Values.garage.s3.web.rootDomain }}"
     index = "{{ .Values.garage.s3.web.index }}"
+    
+    [admin]
+    api_bind_addr = "[::]:3903"
+    {{- if .Values.monitoring.tracing.sink }}
+    trace_sink = "{{ .Values.monitoring.tracing.sink }}"
+    {{- end }}
 
 # Data persistence
 persistence:
@@ -123,6 +129,7 @@ service:
     web:
       port: 3902
     # NOTE: the admin API is excluded for now as it is not consistent across nodes
+
 ingress:
   s3:
     api:
@@ -186,3 +193,23 @@ nodeSelector: {}
 tolerations: []
 
 affinity: {}
+
+monitoring:
+  metrics:
+    # If true, a service for monitoring is created with a prometheus.io/scrape annotation
+    enabled: false
+    serviceMonitor:
+      # If true, a ServiceMonitor CRD is created for a prometheus operator
+      # https://github.com/coreos/prometheus-operator
+      #
+      enabled: false
+      path: /metrics
+      #  namespace: monitoring  (defaults to use the namespace this chart is deployed to)
+      labels: {}
+      interval: 15s
+      scheme: http
+      tlsConfig: {}
+      scrapeTimeout: 10s
+      relabelings: []
+  tracing:
+    sink: ""