fix examples

Author: maltesander

examples/README-examples.md

@@ -50,7 +50,7 @@ Several resources are needed in this store. These can be loaded like this:
 
 ```text
 kubectl exec minio-mc-0 -- sh -c 'mc alias set test-minio http://test-minio:9000/'
-kubectl cp examples/ny-tlc-report-1.1.0-3.5.7.jar  minio-mc-0:/tmp
+kubectl cp tests/templates/kuttl/spark-ny-public-s3/ny-tlc-report-1.1.0-3.5.7.jar  minio-mc-0:/tmp
 kubectl cp apps/ny_tlc_report.py  minio-mc-0:/tmp
 kubectl cp examples/yellow_tripdata_2021-07.csv  minio-mc-0:/tmp
 kubectl exec minio-mc-0 -- mc cp /tmp/ny-tlc-report-1.1.0-3.5.7.jar test-minio/my-bucket

kind/assert-pvc-jars.yaml

@@ -13,7 +13,7 @@ spec:
             claimName: pvc-ksv
       containers:
         - name: assert-pvc-jars
-          image: oci.stackable.tech/sdp/tools:0.2.0-stackable0.4.0
+          image: oci.stackable.tech/sdp/tools:1.0.0-stackable0.0.0-dev
           env:
             - name: DEST_DIR
               value: "/dependencies/jars"

kind/kind-pvc.yaml

@@ -24,7 +24,7 @@ spec:
             claimName: pvc-ksv
       containers:
         - name: aws-deps
-          image: oci.stackable.tech/sdp/tools:0.2.0-stackable0.4.0
+          image: oci.stackable.tech/sdp/tools:1.0.0-stackable0.0.0-dev
           env:
             - name: DEST_DIR
               value: "/dependencies/jars"

kind/minio.yaml

@@ -29,6 +29,6 @@ spec:
     spec:
       containers:
         - name: minio-mc
-          image: bitnamilegacy/minio:2022-debian-10
+          image: docker.io/bitnamilegacy/minio:2024-debian-12
           stdin: true
           tty: true

rust/operator-binary/src/history/history_controller.rs

@@ -39,7 +39,7 @@ use stackable_operator::{
         core::{DeserializeGuard, error_boundary},
         runtime::{controller::Action, reflector::ObjectRef},
     },
-    kvp::{Labels, ObjectLabels},
+    kvp::Labels,
     logging::controller::ReconcilerError,
     product_logging::{
         framework::{LoggingError, calculate_log_volume_size_limit, vector_container},
@@ -57,20 +57,24 @@ use crate::{
     Ctx,
     crd::{
         constants::{
-            ACCESS_KEY_ID, HISTORY_APP_NAME, HISTORY_CONTROLLER_NAME, HISTORY_ROLE_NAME,
-            HISTORY_UI_PORT, JVM_SECURITY_PROPERTIES_FILE, LISTENER_VOLUME_DIR,
-            LISTENER_VOLUME_NAME, MAX_SPARK_LOG_FILES_SIZE, METRICS_PORT, OPERATOR_NAME,
-            SECRET_ACCESS_KEY, SPARK_DEFAULTS_FILE_NAME, SPARK_ENV_SH_FILE_NAME,
-            SPARK_IMAGE_BASE_NAME, STACKABLE_TRUST_STORE, VOLUME_MOUNT_NAME_CONFIG,
-            VOLUME_MOUNT_NAME_LOG, VOLUME_MOUNT_NAME_LOG_CONFIG, VOLUME_MOUNT_PATH_CONFIG,
-            VOLUME_MOUNT_PATH_LOG, VOLUME_MOUNT_PATH_LOG_CONFIG,
+            ACCESS_KEY_ID, HISTORY_APP_NAME, HISTORY_CONTROLLER_NAME, HISTORY_UI_PORT,
+            JVM_SECURITY_PROPERTIES_FILE, LISTENER_VOLUME_DIR, LISTENER_VOLUME_NAME,
+            MAX_SPARK_LOG_FILES_SIZE, METRICS_PORT, OPERATOR_NAME, SECRET_ACCESS_KEY,
+            SPARK_DEFAULTS_FILE_NAME, SPARK_ENV_SH_FILE_NAME, SPARK_IMAGE_BASE_NAME,
+            STACKABLE_TRUST_STORE, VOLUME_MOUNT_NAME_CONFIG, VOLUME_MOUNT_NAME_LOG,
+            VOLUME_MOUNT_NAME_LOG_CONFIG, VOLUME_MOUNT_PATH_CONFIG, VOLUME_MOUNT_PATH_LOG,
+            VOLUME_MOUNT_PATH_LOG_CONFIG,
         },
         history::{self, HistoryConfig, SparkHistoryServerContainer, v1alpha1},
         listener_ext,
         logdir::ResolvedLogDir,
         tlscerts, to_spark_env_sh_string,
     },
-    history::operations::pdb::add_pdbs,
+    history::{
+        operations::pdb::add_pdbs,
+        recommended_labels,
+        service::{self, build_rolegroup_metrics_service},
+    },
     product_logging::{self},
 };
 
@@ -96,9 +100,6 @@ pub enum Error {
     #[snafu(display("missing secret lifetime"))]
     MissingSecretLifetime,
 
-    #[snafu(display("object has no namespace"))]
-    ObjectHasNoNamespace,
-
     #[snafu(display("invalid config map {name}"))]
     InvalidConfigMap {
         source: stackable_operator::builder::configmap::Error,
@@ -125,6 +126,11 @@ pub enum Error {
         source: stackable_operator::cluster_resources::Error,
     },
 
+    #[snafu(display("failed to update history server metrics service"))]
+    ApplyMetricsService {
+        source: stackable_operator::cluster_resources::Error,
+    },
+
     #[snafu(display("failed to apply role ServiceAccount"))]
     ApplyServiceAccount {
         source: stackable_operator::cluster_resources::Error,
@@ -233,6 +239,9 @@ pub enum Error {
     ResolveProductImage {
         source: product_image_selection::Error,
     },
+
+    #[snafu(display("failed to resolve product image"))]
+    BuildMetricsService { source: service::Error },
 }
 
 impl ReconcilerError for Error {
@@ -320,10 +329,10 @@ pub async fn reconcile(
                 &rgr,
                 &log_dir,
             )?;
-            cluster_resources
-                .add(client, config_map)
-                .await
-                .context(ApplyConfigMapSnafu)?;
+
+            let metrics_service =
+                build_rolegroup_metrics_service(shs, &resolved_product_image, &rgr)
+                    .context(BuildMetricsServiceSnafu)?;
 
             let sts = build_stateful_set(
                 shs,
@@ -333,6 +342,15 @@ pub async fn reconcile(
                 &merged_config,
                 &service_account,
             )?;
+
+            cluster_resources
+                .add(client, config_map)
+                .await
+                .context(ApplyConfigMapSnafu)?;
+            cluster_resources
+                .add(client, metrics_service)
+                .await
+                .context(ApplyMetricsServiceSnafu)?;
             cluster_resources
                 .add(client, sts)
                 .await
@@ -380,7 +398,7 @@ fn build_group_listener(
     let listener_name = group_listener_name(shs, role);
 
     let recommended_object_labels =
-        labels(shs, &resolved_product_image.app_version_label_value, "none");
+        recommended_labels(shs, &resolved_product_image.app_version_label_value, "none");
 
     let listener_ports = [listener::v1alpha1::ListenerPort {
         name: "http".to_string(),
@@ -445,7 +463,7 @@ fn build_config_map(
                 .name(&cm_name)
                 .ownerreference_from_resource(shs, None, Some(true))
                 .context(ObjectMissingMetadataForOwnerRefSnafu)?
-                .with_recommended_labels(labels(
+                .with_recommended_labels(recommended_labels(
                     shs,
                     app_version_label_value,
                     &rolegroupref.role_group,
@@ -511,7 +529,7 @@ fn build_stateful_set(
         rolegroupref.object_name()
     };
 
-    let recommended_object_labels = labels(
+    let recommended_object_labels = recommended_labels(
         shs,
         &resolved_product_image.app_version_label_value,
         rolegroupref.role_group.as_ref(),
@@ -732,22 +750,6 @@ fn command_args(logdir: &ResolvedLogDir) -> Vec<String> {
     vec![command.join("\n")]
 }
 
-fn labels<'a, T>(
-    shs: &'a T,
-    app_version_label_value: &'a str,
-    role_group: &'a str,
-) -> ObjectLabels<'a, T> {
-    ObjectLabels {
-        owner: shs,
-        app_name: HISTORY_APP_NAME,
-        app_version: app_version_label_value,
-        operator_name: OPERATOR_NAME,
-        controller_name: HISTORY_CONTROLLER_NAME,
-        role: HISTORY_ROLE_NAME,
-        role_group,
-    }
-}
-
 /// Return the Spark properties for the cleaner role group (if any).
 /// There should be only one role group with "cleaner=true" and this
 /// group should have a replica count of 0 or 1.

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

@@ -1,3 +1,26 @@
+use stackable_operator::kvp::ObjectLabels;
+
+use crate::crd::constants::{
+    HISTORY_APP_NAME, HISTORY_CONTROLLER_NAME, HISTORY_ROLE_NAME, OPERATOR_NAME,
+};
+
 pub mod config;
 pub mod history_controller;
 pub mod operations;
+pub mod service;
+
+pub(crate) fn recommended_labels<'a, T>(
+    shs: &'a T,
+    app_version_label_value: &'a str,
+    role_group: &'a str,
+) -> ObjectLabels<'a, T> {
+    ObjectLabels {
+        owner: shs,
+        app_name: HISTORY_APP_NAME,
+        app_version: app_version_label_value,
+        operator_name: OPERATOR_NAME,
+        controller_name: HISTORY_CONTROLLER_NAME,
+        role: HISTORY_ROLE_NAME,
+        role_group,
+    }
+}

rust/operator-binary/src/history/service.rs

@@ -0,0 +1,105 @@
+use snafu::{ResultExt, Snafu};
+use stackable_operator::{
+    builder::meta::ObjectMetaBuilder,
+    commons::product_image_selection::ResolvedProductImage,
+    k8s_openapi::api::core::v1::{Service, ServicePort, ServiceSpec},
+    kvp::{Annotations, Labels},
+    role_utils::RoleGroupRef,
+};
+
+use crate::{
+    crd::{
+        constants::{HISTORY_APP_NAME, METRICS_PORT},
+        history::v1alpha1,
+    },
+    history::recommended_labels,
+};
+
+#[derive(Snafu, Debug)]
+pub enum Error {
+    #[snafu(display("failed to build Labels"))]
+    LabelBuild {
+        source: stackable_operator::kvp::LabelError,
+    },
+
+    #[snafu(display("failed to build Metadata"))]
+    MetadataBuild {
+        source: stackable_operator::builder::meta::Error,
+    },
+    #[snafu(display("object is missing metadata to build owner reference"))]
+    ObjectMissingMetadataForOwnerRef {
+        source: stackable_operator::builder::meta::Error,
+    },
+}
+
+/// The rolegroup metrics [`Service`] is a service that exposes metrics and a prometheus scraping label
+pub fn build_rolegroup_metrics_service(
+    shs: &v1alpha1::SparkHistoryServer,
+    resolved_product_image: &ResolvedProductImage,
+    rolegroup_ref: &RoleGroupRef<v1alpha1::SparkHistoryServer>,
+) -> Result<Service, Error> {
+    Ok(Service {
+        metadata: ObjectMetaBuilder::new()
+            .name_and_namespace(shs)
+            .name(rolegroup_ref.rolegroup_metrics_service_name())
+            .ownerreference_from_resource(shs, None, Some(true))
+            .context(ObjectMissingMetadataForOwnerRefSnafu)?
+            .with_recommended_labels(recommended_labels(
+                shs,
+                &resolved_product_image.app_version_label_value,
+                &rolegroup_ref.role_group,
+            ))
+            .context(MetadataBuildSnafu)?
+            .with_labels(prometheus_labels())
+            .with_annotations(prometheus_annotations())
+            .build(),
+        spec: Some(ServiceSpec {
+            // Internal communication does not need to be exposed
+            type_: Some("ClusterIP".to_string()),
+            cluster_ip: Some("None".to_string()),
+            ports: Some(metrics_ports()),
+            selector: Some(
+                Labels::role_group_selector(
+                    shs,
+                    HISTORY_APP_NAME,
+                    &rolegroup_ref.role,
+                    &rolegroup_ref.role_group,
+                )
+                .context(LabelBuildSnafu)?
+                .into(),
+            ),
+            publish_not_ready_addresses: Some(true),
+            ..ServiceSpec::default()
+        }),
+        status: None,
+    })
+}
+
+fn metrics_ports() -> Vec<ServicePort> {
+    vec![ServicePort {
+        name: Some("metrics".to_string()),
+        port: METRICS_PORT.into(),
+        protocol: Some("TCP".to_string()),
+        ..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")
+}