feat(data): OpenSearch prometheus-exporter sidecar

elasticsearch-exporter v1.7.0 runs as a sidecar, scrapes localhost:9200,
exposes elasticsearch_* metrics on :9114. ServiceMonitor re-enabled.
Alert rules updated to use elasticsearch_* metric names.
Flags: --es.all --es.indices --es.shards --collector.clustersettings

base/data/kustomization.yaml

@@ -12,9 +12,10 @@ resources:
   - opensearch-service.yaml
   - opensearch-pvc.yaml
   - barman-vault-secret.yaml
-  # opensearch-servicemonitor.yaml removed — OpenSearch 3.x has no prometheus-exporter plugin.
+  - opensearch-servicemonitor.yaml
-  # TODO: add opensearch-exporter sidecar for Prometheus metrics.
   - opensearch-alertrules.yaml
+  - cnpg-podmonitor.yaml
+  - openbao-servicemonitor.yaml
   - postgres-alertrules.yaml
   - openbao-alertrules.yaml
   - searxng-deployment.yaml

base/data/opensearch-alertrules.yaml

@@ -10,7 +10,7 @@ spec:
     - name: opensearch
       rules:
         - alert: OpenSearchClusterRed
-          expr: opensearch_cluster_health_status{color="red"} == 1
+          expr: elasticsearch_cluster_health_status{color="red"} == 1
           for: 2m
           labels:
             severity: critical
@@ -19,7 +19,7 @@ spec:
             description: "OpenSearch cluster {{ $labels.cluster }} health status is red."
 
         - alert: OpenSearchClusterYellow
-          expr: opensearch_cluster_health_status{color="yellow"} == 1
+          expr: elasticsearch_cluster_health_status{color="yellow"} == 1
           for: 10m
           labels:
             severity: warning
@@ -28,10 +28,10 @@ spec:
             description: "OpenSearch cluster {{ $labels.cluster }} health status is yellow."
 
         - alert: OpenSearchHeapHigh
-          expr: (opensearch_jvm_mem_heap_used_bytes / opensearch_jvm_mem_heap_max_bytes) > 0.85
+          expr: (elasticsearch_jvm_memory_used_bytes{area="heap"} / elasticsearch_jvm_memory_max_bytes{area="heap"}) > 0.85
           for: 5m
           labels:
             severity: warning
           annotations:
             summary: "OpenSearch JVM heap usage is high"
-            description: "OpenSearch node {{ $labels.node }} in {{ $labels.namespace }} heap usage is above 85%."
+            description: "OpenSearch node {{ $labels.name }} in {{ $labels.namespace }} heap usage is above 85%."

base/data/opensearch-deployment.yaml

@@ -24,9 +24,6 @@ spec:
       containers:
         - name: opensearch
           image: opensearchproject/opensearch:3
-          # OpenSearch 3.x has no maintained prometheus-exporter plugin.
-          # Metrics come from /_cluster/stats JSON API (scraped by dashboard queries).
-          # TODO: add opensearch-exporter sidecar for native Prometheus metrics.
           ports:
             - name: http
               containerPort: 9200
@@ -58,6 +55,26 @@ spec:
           volumeMounts:
             - name: data
               mountPath: /usr/share/opensearch/data
+        # Prometheus metrics exporter — scrapes OpenSearch REST APIs and
+        # exposes them as elasticsearch_* metrics on :9114/metrics.
+        - name: exporter
+          image: quay.io/prometheuscommunity/elasticsearch-exporter:v1.7.0
+          args:
+            - --es.uri=http://localhost:9200
+            - --es.all
+            - --es.indices
+            - --es.shards
+            - --collector.clustersettings
+          ports:
+            - name: metrics
+              containerPort: 9114
+              protocol: TCP
+          resources:
+            requests:
+              cpu: 5m
+              memory: 32Mi
+            limits:
+              memory: 64Mi
       volumes:
         - name: data
           persistentVolumeClaim:

base/data/opensearch-service.yaml

@@ -17,3 +17,7 @@ spec:
       port: 9300
       targetPort: 9300
       protocol: TCP
+    - name: metrics
+      port: 9114
+      targetPort: 9114
+      protocol: TCP

base/data/opensearch-servicemonitor.yaml

@@ -11,6 +11,5 @@ spec:
     matchLabels:
       app: opensearch
   endpoints:
-    - port: http
+    - port: metrics
       interval: 30s
-      path: /_prometheus/metrics