cleanup obsolete service monitors

Author: maltesander

stacks/monitoring/prometheus-service-monitors.yaml

@@ -1,72 +1,21 @@
+---
 # Use something like this to check for metrics:
 # count by (app_kubernetes_io_name, app_kubernetes_io_instance, pod) ({app_kubernetes_io_name!="",pod!=""})
 #
 # Products metrics covered by the ServiceMonitors below. The list also includes whether the
 # ServiceMonitor scrapes native metrics or a statsd/JMX exporter.
 #
-# [x] Airflow - exporter
-# [x] Druid - native
-# [x] HBase - native
-# [x] Hadoop HDFS - native
-# [x] Hive - exporter
-# [x] Kafka - exporter
-# [x] NiFi 1 - native
-# [x] NiFi 2 - native
-# [x] OpenSearch - native
-# [ ] Spark - native - partially done, see comment on it below
-# [x] Superset - exporter
-# [x] Trino - native
-# [x] ZooKeeper - native
-# [x] OPA - native
----
-apiVersion: monitoring.coreos.com/v1
-kind: ServiceMonitor
-metadata:
-  name: stackable
-  labels:
-    stackable.tech/vendor: Stackable
-    release: prometheus
-spec:
-  namespaceSelector:
-    any: true
-  selector:
-    matchLabels:
-      stackable.tech/vendor: Stackable
-      prometheus.io/scrape: "true"
-    matchExpressions:
-      - key: app.kubernetes.io/name
-        operator: In
-        values:
-          - airflow
-          - druid
-          - hive
-          - kafka
-          - nifi # This only works for NiFi 1, NiFi 2 works via stackable-generic
-          - opa
-          - superset
-          - trino
-          - zookeeper
-  endpoints:
-    - scheme: http
-      port: metrics
-      path: /metrics
-  podTargetLabels:
-    - app.kubernetes.io/name
-    - app.kubernetes.io/instance
-    - app.kubernetes.io/component
-    - app.kubernetes.io/role-group
-    - app.kubernetes.io/version
----
+#
 # Utilize `prometheus.io/scheme`, `prometheus.io/port`, `prometheus.io/path` annotations set by the operators
 # to scrape all Stackable products.
 # [x] Airflow - relabel drop filter on airflow container
 # [x] Druid
 # [x] HBase
 # [X] Hadoop HDFS - relabel drop filter on empty container
 # [x] Hive
-# [~] Kafka - TODO: listener services have metrics?
+# [x] Kafka - TODO: listener services have metrics?
 # [x] NiFi 1 + 2
-# [ ] OpenSearch
+# [x] OpenSearch
 # [x] Spark: Connect, HistoryServer
 # [x] Superset - relabel drop filter on superset container
 # [x] Trino
@@ -75,7 +24,7 @@ spec:
 apiVersion: monitoring.coreos.com/v1
 kind: ServiceMonitor
 metadata:
-  name: stackable-generic
+  name: stackable
   labels:
     stackable.tech/vendor: Stackable
     release: prometheus
@@ -91,8 +40,7 @@ spec:
         - sourceLabels:
             - __meta_kubernetes_pod_container_name
           # Pods show up twice due to multiple containers, we only keep the main / product container.
-          # Except for Airflow and Superset, where we chose the metrics container (otherwise scheduler, worker etc.
-          # which only have the metrics container are not getting picked up).
+          # Except for Airflow and Superset, where we chose the metrics container.
           # - airflow: airflow
           # - superset: superset
           # - empty: filter when container label does not exist: hdfs
@@ -136,109 +84,6 @@ spec:
     - app.kubernetes.io/role-group
     - app.kubernetes.io/version
 ---
-# We prefer the native metrics over the statsd-exporter
-apiVersion: monitoring.coreos.com/v1
-kind: ServiceMonitor
-metadata:
-  name: stackable-hdfs
-  labels:
-    stackable.tech/vendor: Stackable
-    release: prometheus
-spec:
-  namespaceSelector:
-    any: true
-  selector:
-    matchLabels:
-      stackable.tech/vendor: Stackable
-      prometheus.io/scrape: "true"
-      app.kubernetes.io/name: hdfs
-  endpoints:
-    - scheme: http
-      port: http # Don't use the "metrics" exporter port, we want native metrics instead
-      path: /prom
-  podTargetLabels:
-    - app.kubernetes.io/name
-    - app.kubernetes.io/instance
-    - app.kubernetes.io/component
-    - app.kubernetes.io/role-group
-    - app.kubernetes.io/version
----
-apiVersion: monitoring.coreos.com/v1
-kind: ServiceMonitor
-metadata:
-  name: stackable-hbase
-  labels:
-    stackable.tech/vendor: Stackable
-    release: prometheus
-spec:
-  namespaceSelector:
-    any: true
-  selector:
-    matchLabels:
-      stackable.tech/vendor: Stackable
-      prometheus.io/scrape: "true"
-      app.kubernetes.io/name: hbase
-  endpoints:
-    - scheme: http
-      port: metrics
-      path: /prometheus
-  podTargetLabels:
-    - app.kubernetes.io/name
-    - app.kubernetes.io/instance
-    - app.kubernetes.io/component
-    - app.kubernetes.io/role-group
-    - app.kubernetes.io/version
----
-apiVersion: monitoring.coreos.com/v1
-kind: ServiceMonitor
-metadata:
-  name: stackable-opensearch
-  labels:
-    stackable.tech/vendor: Stackable
-    release: prometheus
-spec:
-  namespaceSelector:
-    any: true
-  selector:
-    matchLabels:
-      stackable.tech/vendor: Stackable
-      prometheus.io/scrape: "true"
-      app.kubernetes.io/name: opensearch
-  endpoints:
-    - relabelings:
-        - sourceLabels:
-            - __meta_kubernetes_service_annotation_prometheus_io_scheme
-          action: replace
-          targetLabel: __scheme__
-          regex: (https?)
-        - sourceLabels:
-            - __meta_kubernetes_service_annotation_prometheus_io_path
-          action: replace
-          targetLabel: __metrics_path__
-          regex: (.+)
-        # Use the FQDN instead of the IP address because the IP address
-        # is not contained in the certificate.
-        - sourceLabels:
-            - __meta_kubernetes_pod_name
-            - __meta_kubernetes_service_name
-            - __meta_kubernetes_namespace
-            - __meta_kubernetes_service_annotation_prometheus_io_port
-          action: replace
-          targetLabel: __address__
-          regex: (.+);(.+);(.+);(\d+)
-          replacement: $1.$2.$3.svc.cluster.local:$4
-      tlsConfig:
-        ca:
-          secret:
-            name: prometheus-tls-certificate
-            key: ca.crt
-  podTargetLabels:
-    - app.kubernetes.io/name
-    - app.kubernetes.io/instance
-    - app.kubernetes.io/component
-    - app.kubernetes.io/role-group
-    - app.kubernetes.io/version
----
 # spark-k8s-operator does not deploy any Services at all (at least for SparkApplications).
 # We currently only scrape the driver, going forward we might want to scrape the executors as well.
 # In the future we might also want to scrape SparkConnect and HistoryServers.
@@ -271,26 +116,6 @@ spec:
 ---
 apiVersion: monitoring.coreos.com/v1
 kind: ServiceMonitor
-metadata:
-  name: stackable-minio-http
-  labels:
-    stackable.tech/vendor: Stackable
-    release: prometheus
-spec:
-  namespaceSelector:
-    any: true
-  selector:
-    matchLabels:
-      # stackable.tech/vendor: Stackable # This is not always set, e.g. missing in the nifi-kafka-druid-water-level-data demo
-      app: minio
-      monitoring: "true"
-  endpoints:
-    - scheme: http
-      port: http
-      path: /minio/v2/metrics/cluster
----
-apiVersion: monitoring.coreos.com/v1
-kind: ServiceMonitor
 metadata:
   name: stackable-minio-https
   labels: