[Merged by Bors] - Implement resource requests and limits for Superset pods

CHANGELOG.md

@@ -2,6 +2,12 @@
 
 ## [Unreleased]
 
+### Added
+
+- CPU and memory limits are now configurable ([#273]).
+
+[#273]: https://github.com/stackabletech/superset-operator/pull/273
+
 ## [0.6.0] - 2022-09-07
 
 ### Added

docs/modules/ROOT/pages/usage.adoc

@@ -166,3 +166,64 @@ nodes:
 ----
 
 // cliOverrides don't make sense for this operator, so the feature is omitted for now
+
+=== Storage for data volumes
+
+The Superset operator currently does not support using https://kubernetes.io/docs/concepts/storage/persistent-volumes[PersistentVolumeClaims] for internal storage.
+
+=== Memory requests
+
+You can request a certain amount of memory for each individual role group as shown below:
+
+[source,yaml]
+----
+nodes:
+  roleGroups:
+    default:
+      config:
+        resources:
+          memory:
+            limit: '2Gi'
+----
+
+In this example, each Superset container in the `default` role group will have a maximum of 2 gigabytes of memory. To be more precise, these memory limits apply to the container running Superset but not to any sidecar containers (e.g. the metrics container) that are part of the pod.
+
+For more details regarding Kubernetes memory requests and limits see: https://kubernetes.io/docs/tasks/configure-pod-container/assign-memory-resource/[Assign Memory Resources to Containers and Pods].
+
+=== CPU requests
+
+Similarly to memory resources, you can also configure CPU limits, as shown below:
+
+[source,yaml]
+----
+nodes:
+  roleGroups:
+    default:
+      config:
+        resources:
+          cpu:
+            max: '500m'
+            min: '250m'
+----
+
+=== Defaults
+
+If nothing is specified, the operator will automatically set the following default values for resources:
+
+[source,yaml]
+----
+nodes:
+  roleGroups:
+    default:
+      config:
+        resources:
+          cpu:
+            min: '200m'
+            max: "4"
+          memory:
+            limit: '2Gi'
+----
+
+WARNING: The default values are _most likely_ not sufficient to run a proper cluster in production. Please adapt according to your requirements.
+
+For more details regarding Kubernetes CPU limits see: https://kubernetes.io/docs/tasks/configure-pod-container/assign-cpu-resource/[Assign CPU Resources to Containers and Pods].

rust/crd/Cargo.toml

@@ -13,3 +13,4 @@ serde_json = "1.0"
 stackable-operator = { git = "https://github.com/stackabletech/operator-rs.git", tag = "0.25.0" }
 strum = { version = "0.24", features = ["derive"] }
 snafu = "0.7"
+tracing = "0.1"

rust/crd/src/lib.rs

@@ -1,19 +1,19 @@
 pub mod druidconnection;
 pub mod supersetdb;
 
-use std::collections::BTreeMap;
-use std::num::ParseIntError;
-
 use serde::{Deserialize, Serialize};
 use snafu::Snafu;
-use stackable_operator::kube::runtime::reflector::ObjectRef;
-use stackable_operator::kube::CustomResource;
-use stackable_operator::product_config::flask_app_config_writer::{
-    FlaskAppConfigOptions, PythonType,
+use stackable_operator::{
+    commons::resources::{CpuLimits, MemoryLimits, NoRuntimeLimits, Resources},
+    config::merge::Merge,
+    k8s_openapi::apimachinery::pkg::api::resource::Quantity,
+    kube::{runtime::reflector::ObjectRef, CustomResource},
+    product_config::flask_app_config_writer::{FlaskAppConfigOptions, PythonType},
+    product_config_utils::{ConfigError, Configuration},
+    role_utils::{Role, RoleGroupRef},
+    schemars::{self, JsonSchema},
 };
-use stackable_operator::product_config_utils::{ConfigError, Configuration};
-use stackable_operator::role_utils::{Role, RoleGroupRef};
-use stackable_operator::schemars::{self, JsonSchema};
+use std::{collections::BTreeMap, num::ParseIntError};
 use strum::{Display, EnumIter, EnumString, IntoEnumIterator};
 
 pub const APP_NAME: &str = "superset";
@@ -236,7 +236,11 @@ pub enum SupersetRole {
     Node,
 }
 
-#[derive(Clone, Debug, Default, Deserialize, Eq, JsonSchema, PartialEq, Serialize)]
+#[derive(Clone, Debug, Default, Deserialize, Eq, Merge, JsonSchema, PartialEq, Serialize)]
+#[serde(rename_all = "camelCase")]
+pub struct SupersetStorageConfig {}
+
+#[derive(Clone, Debug, Default, Deserialize, JsonSchema, PartialEq, Serialize)]
 #[serde(rename_all = "camelCase")]
 pub struct SupersetConfig {
     /// Row limit when requesting chart data.
@@ -247,11 +251,27 @@ pub struct SupersetConfig {
     /// If you get timeout errors before your query returns the result you may need to increase this timeout.
     /// Corresponds to SUPERSET_WEBSERVER_TIMEOUT
     pub webserver_timeout: Option<u32>,
+    /// CPU and memory limits for Superset pods
+    pub resources: Option<Resources<SupersetStorageConfig, NoRuntimeLimits>>,
 }
 
 impl SupersetConfig {
     pub const CREDENTIALS_SECRET_PROPERTY: &'static str = "credentialsSecret";
     pub const MAPBOX_SECRET_PROPERTY: &'static str = "mapboxSecret";
+
+    fn default_resources() -> Resources<SupersetStorageConfig, NoRuntimeLimits> {
+        Resources {
+            cpu: CpuLimits {
+                min: Some(Quantity("200m".to_owned())),
+                max: Some(Quantity("4".to_owned())),
+            },
+            memory: MemoryLimits {
+                limit: Some(Quantity("2Gi".to_owned())),
+                runtime_limits: NoRuntimeLimits {},
+            },
+            storage: SupersetStorageConfig {},
+        }
+    }
 }
 
 impl Configuration for SupersetConfig {
@@ -323,6 +343,42 @@ impl SupersetCluster {
             role_group: group_name.into(),
         }
     }
+
+    /// Retrieve and merge resource configs for role and role groups
+    pub fn resolve_resource_config_for_role_and_rolegroup(
+        &self,
+        role: &SupersetRole,
+        rolegroup_ref: &RoleGroupRef<SupersetCluster>,
+    ) -> Option<Resources<SupersetStorageConfig, NoRuntimeLimits>> {
+        // Initialize the result with all default values as baseline
+        let conf_defaults = SupersetConfig::default_resources();
+
+        let role = match role {
+            SupersetRole::Node => self.spec.nodes.as_ref()?,
+        };
+
+        // Retrieve role resource config
+        let mut conf_role: Resources<SupersetStorageConfig, NoRuntimeLimits> =
+            role.config.config.resources.clone().unwrap_or_default();
+
+        // Retrieve rolegroup specific resource config
+        let mut conf_rolegroup: Resources<SupersetStorageConfig, NoRuntimeLimits> = role
+            .role_groups
+            .get(&rolegroup_ref.role_group)
+            .and_then(|rg| rg.config.config.resources.clone())
+            .unwrap_or_default();
+
+        // Merge more specific configs into default config
+        // Hierarchy is:
+        // 1. RoleGroup
+        // 2. Role
+        // 3. Default
+        conf_role.merge(&conf_defaults);
+        conf_rolegroup.merge(&conf_role);
+
+        tracing::debug!("Merged resource config: {:?}", conf_rolegroup);
+        Some(conf_rolegroup)
+    }
 }
 
 /// A reference to a [`SupersetCluster`]

rust/operator-binary/src/superset_controller.rs

@@ -15,6 +15,7 @@ use stackable_operator::{
     cluster_resources::ClusterResources,
     commons::{
         authentication::{AuthenticationClass, AuthenticationClassProvider},
+        resources::{NoRuntimeLimits, Resources},
         secret_class::SecretClassVolumeScope,
         tls::{CaCert, TlsServerVerification, TlsVerification},
     },
@@ -43,8 +44,8 @@ use stackable_operator::{
 };
 use stackable_superset_crd::{
     supersetdb::{SupersetDB, SupersetDBStatusCondition},
-    SupersetCluster, SupersetConfig, SupersetConfigOptions, SupersetRole, PYTHONPATH,
-    SUPERSET_CONFIG_FILENAME,
+    SupersetCluster, SupersetConfig, SupersetConfigOptions, SupersetRole, SupersetStorageConfig,
+    PYTHONPATH, SUPERSET_CONFIG_FILENAME,
 };
 use std::{
     borrow::Cow,
@@ -160,6 +161,8 @@ pub enum Error {
     MissingSupersetConfigInNodeConfig,
     #[snafu(display("failed to get {timeout} from {SUPERSET_CONFIG_FILENAME} file. It should be set in the product config or by user input", timeout = SupersetConfigOptions::SupersetWebserverTimeout))]
     MissingWebServerTimeoutInSupersetConfig,
+    #[snafu(display("failed to resolve and merge resource config for role and role group"))]
+    FailedToResolveResourceConfig,
 }
 
 type Result<T, E = Error> = std::result::Result<T, E>;
@@ -269,6 +272,10 @@ pub async fn reconcile_superset(superset: Arc<SupersetCluster>, ctx: Arc<Ctx>) -
     for (rolegroup_name, rolegroup_config) in role_node_config.iter() {
         let rolegroup = superset.node_rolegroup_ref(rolegroup_name);
 
+        let resources = superset
+            .resolve_resource_config_for_role_and_rolegroup(&SupersetRole::Node, &rolegroup)
+            .context(FailedToResolveResourceConfigSnafu)?;
+
         let rg_service = build_node_rolegroup_service(&rolegroup, &superset)?;
         let rg_configmap = build_rolegroup_config_map(
             &superset,
@@ -281,6 +288,7 @@ pub async fn reconcile_superset(superset: Arc<SupersetCluster>, ctx: Arc<Ctx>) -
             &superset,
             rolegroup_config,
             authentication_class.as_ref(),
+            &resources,
         )?;
         cluster_resources
             .add(client, &rg_service)
@@ -479,6 +487,7 @@ fn build_server_rolegroup_statefulset(
     superset: &SupersetCluster,
     node_config: &HashMap<PropertyNameKind, BTreeMap<String, String>>,
     authentication_class: Option<&AuthenticationClass>,
+    resources: &Resources<SupersetStorageConfig, NoRuntimeLimits>,
 ) -> Result<StatefulSet> {
     let rolegroup = superset
         .spec
@@ -551,6 +560,7 @@ fn build_server_rolegroup_statefulset(
                 'superset.app:create_app()'
             "},
         ])
+        .resources(resources.clone().into())
         .build();
     let metrics_container = ContainerBuilder::new("metrics")
         .expect("ContainerBuilder not created")

tests/templates/kuttl/resources/00-assert.yaml

@@ -0,0 +1,14 @@
+---
+apiVersion: kuttl.dev/v1beta1
+kind: TestAssert
+metadata:
+  name: test-superset-postgresql
+timeout: 480
+---
+apiVersion: apps/v1
+kind: StatefulSet
+metadata:
+  name: superset-postgresql
+status:
+  readyReplicas: 1
+  replicas: 1

tests/templates/kuttl/resources/00-install-postgresql.yaml

@@ -0,0 +1,12 @@
+---
+apiVersion: kuttl.dev/v1beta1
+kind: TestStep
+commands:
+  - script: >-
+      helm install superset-postgresql
+      --namespace $NAMESPACE
+      --version 11.0.0
+      --set auth.username=superset
+      --set auth.password=superset
+      --set auth.database=superset
+      --repo https://charts.bitnami.com/bitnami postgresql

tests/templates/kuttl/resources/01-assert.yaml

@@ -0,0 +1,48 @@
+---
+apiVersion: kuttl.dev/v1beta1
+kind: TestAssert
+metadata:
+  name: install-superset
+timeout: 300
+---
+apiVersion: apps/v1
+kind: StatefulSet
+metadata:
+  name: superset-node-resources-from-role
+spec:
+  template:
+    spec:
+      containers:
+        - name: superset
+          resources:
+            requests:
+              cpu: 100m
+              memory: 1Gi
+            limits:
+              cpu: "1"
+              memory: 1Gi
+        - name: metrics
+status:
+  readyReplicas: 1
+  replicas: 1
+---
+apiVersion: apps/v1
+kind: StatefulSet
+metadata:
+  name: superset-node-resources-from-role-group
+spec:
+  template:
+    spec:
+      containers:
+        - name: superset
+          resources:
+            requests:
+              cpu: 300m
+              memory: 3Gi
+            limits:
+              cpu: "3"
+              memory: 3Gi
+        - name: metrics
+status:
+  readyReplicas: 1
+  replicas: 1

tests/templates/kuttl/resources/01-install-superset.yaml.j2

@@ -0,0 +1,48 @@
+---
+apiVersion: kuttl.dev/v1beta1
+kind: TestStep
+timeout: 300
+---
+apiVersion: v1
+kind: Secret
+metadata:
+  name: superset-credentials
+type: Opaque
+stringData:
+  adminUser.username: admin
+  adminUser.firstname: Superset
+  adminUser.lastname: Admin
+  adminUser.email: admin@superset.com
+  adminUser.password: admin
+  connections.secretKey: thisISaSECRET_1234
+  connections.sqlalchemyDatabaseUri: postgresql://superset:superset@superset-postgresql/superset
+---
+apiVersion: superset.stackable.tech/v1alpha1
+kind: SupersetCluster
+metadata:
+  name: superset
+spec:
+  version: {{ test_scenario['values']['superset-latest'] }}
+  statsdExporterVersion: v0.22.4
+  credentialsSecret: superset-credentials
+  loadExamplesOnInit: false
+  nodes:
+    config:
+      resources:
+        cpu:
+          min: 100m
+          max: "1"
+        memory:
+          limit: 1Gi
+    roleGroups:
+      resources-from-role:
+        replicas: 1
+      resources-from-role-group:
+        replicas: 1
+        config:
+          resources:
+            cpu:
+              min: 300m
+              max: "3"
+            memory:
+              limit: 3Gi