chore: ensure metrics are correctly exposed (#807)

* add prometheus annotations to metrics service

* adapted changelog

Author: maltesander

CHANGELOG.md

@@ -11,6 +11,7 @@ All notable changes to this project will be documented in this file.
 - Helm: Allow Pod `priorityClassName` to be configured ([#798]).
 - Add support for Trino 477 ([#801]).
 - Add support for Hive 4.1.0 ([#805]).
+- Add `prometheus.io/path|port|scheme` annotations to metrics service ([#807]).
 
 ### Changed
 
@@ -36,6 +37,7 @@ All notable changes to this project will be documented in this file.
 [#798]: https://github.com/stackabletech/trino-operator/pull/798
 [#801]: https://github.com/stackabletech/trino-operator/pull/801
 [#805]: https://github.com/stackabletech/trino-operator/pull/805
+[#807]: https://github.com/stackabletech/trino-operator/pull/807
 
 ## [25.7.0] - 2025-07-23
 

rust/operator-binary/src/controller.rs

@@ -92,7 +92,7 @@ use crate::{
         authentication::resolve_authentication_classes,
         catalog,
         discovery::{TrinoDiscovery, TrinoDiscoveryProtocol, TrinoPodRef},
-        rolegroup_headless_service_name, v1alpha1,
+        v1alpha1,
     },
     listener::{
         LISTENER_VOLUME_DIR, LISTENER_VOLUME_NAME, build_group_listener, build_group_listener_pvc,
@@ -156,11 +156,6 @@ pub enum Error {
         source: stackable_operator::cluster_resources::Error,
     },
 
-    #[snafu(display("failed to apply global Service"))]
-    ApplyRoleService {
-        source: stackable_operator::cluster_resources::Error,
-    },
-
     #[snafu(display("failed to apply Service for {}", rolegroup))]
     ApplyRoleGroupService {
         source: stackable_operator::cluster_resources::Error,
@@ -1394,9 +1389,7 @@ fn build_rolegroup_statefulset(
                 ),
                 ..LabelSelector::default()
             },
-            service_name: Some(rolegroup_headless_service_name(
-                &role_group_ref.object_name(),
-            )),
+            service_name: Some(role_group_ref.rolegroup_headless_service_name()),
             template: pod_template,
             volume_claim_templates: Some(persistent_volume_claims),
             ..StatefulSetSpec::default()

rust/operator-binary/src/crd/mod.rs

@@ -120,9 +120,6 @@ pub const MAX_TRINO_LOG_FILES_SIZE: MemoryQuantity = MemoryQuantity {
     unit: BinaryMultiple::Mebi,
 };
 
-pub const METRICS_SERVICE_SUFFIX: &str = "metrics";
-pub const HEADLESS_SERVICE_SUFFIX: &str = "headless";
-
 pub const JVM_HEAP_FACTOR: f32 = 0.8;
 
 pub const DEFAULT_COORDINATOR_GRACEFUL_SHUTDOWN_TIMEOUT: Duration =
@@ -860,9 +857,7 @@ impl v1alpha1::TrinoCluster {
                 let ns = ns.clone();
                 (0..rolegroup.replicas.unwrap_or(0)).map(move |i| TrinoPodRef {
                     namespace: ns.clone(),
-                    role_group_service_name: rolegroup_headless_service_name(
-                        &role_group_ref.object_name(),
-                    ),
+                    role_group_service_name: role_group_ref.rolegroup_headless_service_name(),
                     pod_name: format!(
                         "{role_group}-{i}",
                         role_group = role_group_ref.object_name()
@@ -963,16 +958,6 @@ impl v1alpha1::TrinoCluster {
     }
 }
 
-/// Returns the metrics rolegroup service name `<cluster>-<role>-<rolegroup>-<METRICS_SERVICE_SUFFIX>`.
-pub fn rolegroup_metrics_service_name(role_group_ref_object_name: &str) -> String {
-    format!("{role_group_ref_object_name}-{METRICS_SERVICE_SUFFIX}")
-}
-
-/// Returns the headless rolegroup service name `<cluster>-<role>-<rolegroup>-<HEADLESS_SERVICE_SUFFIX>`.
-pub fn rolegroup_headless_service_name(role_group_ref_object_name: &str) -> String {
-    format!("{role_group_ref_object_name}-{HEADLESS_SERVICE_SUFFIX}")
-}
-
 fn extract_role_from_coordinator_config(
     fragment: Role<TrinoConfigFragment, TrinoCoordinatorRoleConfig, JavaCommonConfig>,
 ) -> Role<TrinoConfigFragment, GenericRoleConfig, JavaCommonConfig> {

rust/operator-binary/src/service.rs

@@ -4,14 +4,11 @@ use snafu::{ResultExt, Snafu};
 use stackable_operator::{
     builder::meta::ObjectMetaBuilder,
     k8s_openapi::api::core::v1::{Service, ServicePort, ServiceSpec},
-    kvp::{Label, ObjectLabels},
+    kvp::{Annotations, Labels, ObjectLabels},
     role_utils::RoleGroupRef,
 };
 
-use crate::crd::{
-    METRICS_PORT, METRICS_PORT_NAME, rolegroup_headless_service_name,
-    rolegroup_metrics_service_name, v1alpha1,
-};
+use crate::crd::{METRICS_PORT, METRICS_PORT_NAME, v1alpha1};
 
 #[derive(Snafu, Debug)]
 pub enum Error {
@@ -42,9 +39,7 @@ pub fn build_rolegroup_headless_service(
     Ok(Service {
         metadata: ObjectMetaBuilder::new()
             .name_and_namespace(trino)
-            .name(rolegroup_headless_service_name(
-                &role_group_ref.object_name(),
-            ))
+            .name(role_group_ref.rolegroup_headless_service_name())
             .ownerreference_from_resource(trino, None, Some(true))
             .context(ObjectMissingMetadataForOwnerRefSnafu)?
             .with_recommended_labels(object_labels)
@@ -73,14 +68,13 @@ pub fn build_rolegroup_metrics_service(
     Ok(Service {
         metadata: ObjectMetaBuilder::new()
             .name_and_namespace(trino)
-            .name(rolegroup_metrics_service_name(
-                &role_group_ref.object_name(),
-            ))
+            .name(role_group_ref.rolegroup_metrics_service_name())
             .ownerreference_from_resource(trino, None, Some(true))
             .context(ObjectMissingMetadataForOwnerRefSnafu)?
             .with_recommended_labels(object_labels)
             .context(MetadataBuildSnafu)?
-            .with_label(Label::try_from(("prometheus.io/scrape", "true")).context(LabelBuildSnafu)?)
+            .with_labels(prometheus_labels())
+            .with_annotations(prometheus_annotations())
             .build(),
         spec: Some(ServiceSpec {
             // Internal communication does not need to be exposed
@@ -115,3 +109,23 @@ fn metrics_service_ports() -> Vec<ServicePort> {
         ..ServicePort::default()
     }]
 }
+
+/// Common labels for Prometheus
+fn prometheus_labels() -> Labels {
+    Labels::try_from([("prometheus.io/scrape", "true")]).expect("should be a valid label")
+}
+
+/// Common annotations for Prometheus
+///
+/// These annotations can be used in a ServiceMonitor.
+///
+/// see also <https://github.com/prometheus-community/helm-charts/blob/prometheus-27.32.0/charts/prometheus/values.yaml#L983-L1036>
+fn prometheus_annotations() -> Annotations {
+    Annotations::try_from([
+        ("prometheus.io/path".to_owned(), "/metrics".to_owned()),
+        ("prometheus.io/port".to_owned(), METRICS_PORT.to_string()),
+        ("prometheus.io/scheme".to_owned(), "http".to_owned()),
+        ("prometheus.io/scrape".to_owned(), "true".to_owned()),
+    ])
+    .expect("should be valid annotations")
+}