• Added PrometheusOperatorAdmissionWebhookConfig struct to ClusterMonitoringSpec with support
 for compute resource requests/limits and pod topology spread constraints
• Defined new configuration type with Resources and TopologySpreadConstraints fields, including
 comprehensive kubebuilder validation markers
• Generated all required Go code artifacts: OpenAPI schemas, Swagger documentation, and deepcopy
 methods
• Updated all CRD manifests (DevPreviewNoUpgrade, CustomNoUpgrade, TechPreviewNoUpgrade, and
 featuregated variants) across both config and payload directories with the new admission webhook
 configuration schema
• Added 10 comprehensive validation test cases covering valid configurations, edge cases, and error
 scenarios for the new admission webhook config

Add OpenAPI schema for Prometheus Operator admission webhook config

• Added OpenAPI schema definition for PrometheusOperatorAdmissionWebhookConfig type
• Registered the new schema in the OpenAPI definitions map
• Updated ClusterMonitoringSpec schema to include prometheusOperatorAdmissionWebhookConfig field
 with detailed description
• Added PrometheusOperatorAdmissionWebhookConfig to the dependencies list for
 ClusterMonitoringSpec

openapi/generated_openapi/zz_generated.openapi.go

Add Swagger documentation for admission webhook configuration

• Added swagger documentation map for PrometheusOperatorAdmissionWebhookConfig struct
• Documented the resources field describing compute resource requests and limits
• Documented the topologySpreadConstraints field for pod distribution across topology domains
• Added field documentation to ClusterMonitoringSpec for the new admission webhook config

config/v1alpha1/zz_generated.swagger_doc_generated.go

Define PrometheusOperatorAdmissionWebhookConfig type and add to spec

• Added PrometheusOperatorAdmissionWebhookConfig field to ClusterMonitoringSpec struct
• Defined new PrometheusOperatorAdmissionWebhookConfig struct with Resources and
 TopologySpreadConstraints fields
• Added comprehensive comments documenting the admission webhook component and its configuration
 options
• Applied kubebuilder validation markers for field constraints (MinProperties, MaxItems, MinItems,
 list types)

config/v1alpha1/types_cluster_monitoring.go

Description

The new PrometheusOperatorAdmissionWebhookConfig TechPreview API type is missing the
+openshift:enable:FeatureGate=<FeatureGateName> marker on both the new field in
ClusterMonitoringSpec and the struct definition itself. No corresponding FeatureGate entry for
this type was added to features/features.go, violating the requirement that every new TechPreview
API type be explicitly gated.

Code

config/v1alpha1/types_cluster_monitoring.go[R518-523]

+// PrometheusOperatorAdmissionWebhookConfig provides configuration options for the admission webhook
+// component of Prometheus Operator that runs in the `openshift-monitoring` namespace. The admission
+// webhook validates PrometheusRule and AlertmanagerConfig objects, mutates PrometheusRule annotations,
+// and converts AlertmanagerConfig objects between API versions.
+// +kubebuilder:validation:MinProperties=1
+type PrometheusOperatorAdmissionWebhookConfig struct {

Evidence

PR Compliance ID 10 requires every new TechPreview API type to have a corresponding FeatureGate
entry in features/features.go and the +openshift:enable:FeatureGate=<FeatureGateName> marker.
The new PrometheusOperatorAdmissionWebhookConfig struct has neither marker nor a new FeatureGate
in features/features.go. The only FeatureGate marker in types_cluster_monitoring.go is on the
top-level ClusterMonitoring type (+openshift:enable:FeatureGate=ClusterMonitoringConfig), and
features/features.go only defines FeatureGateClusterMonitoringConfig with no new entry for this
webhook config type.

AGENTS.md
config/v1alpha1/types_cluster_monitoring.go[518-523]
config/v1alpha1/types_cluster_monitoring.go[102-109]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution
## Issue description
The new `PrometheusOperatorAdmissionWebhookConfig` TechPreview API type and its corresponding field on `ClusterMonitoringSpec` are missing the `+openshift:enable:FeatureGate=<FeatureGateName>` marker. No new FeatureGate entry was added to `features/features.go` for this type.
## Issue Context
Per the project's TechPreview API policy (PR Compliance ID 10), every new TechPreview API type must be gated by a FeatureGate. The parent `ClusterMonitoring` type uses `+openshift:enable:FeatureGate=ClusterMonitoringConfig` (defined in `features/features.go`). If this new type is intended to be included under the same gate, add the marker explicitly to both the field and the struct. If a new separate gate is required, add a new entry to `features/features.go` and reference it in the markers.
## Fix Focus Areas
- config/v1alpha1/types_cluster_monitoring.go[102-109]
- config/v1alpha1/types_cluster_monitoring.go[518-523]
- features/features.go[524-525]

i Copy this prompt and use it to remediate the issue with your preferred AI generation tools

Description

PrometheusOperatorAdmissionWebhookConfig omits nodeSelector and tolerations fields that are
present in both PrometheusOperatorConfig and MetricsServerConfig. Without tolerations, the admission
webhook pod cannot be scheduled on nodes with required taints, potentially making it unschedulable
and blocking all PrometheusRule and AlertmanagerConfig operations.

Code

config/v1alpha1/types_cluster_monitoring.go[R523-559]

+type PrometheusOperatorAdmissionWebhookConfig struct {
+	// resources defines the compute resource requests and limits for the
+	// prometheus-operator-admission-webhook container.
+	// This includes CPU, memory and HugePages constraints to help control scheduling and resource usage.
+	// When not specified, defaults are used by the platform. Requests cannot exceed limits.
+	// This field is optional.
+	// More info: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/
+	// This is a simplified API that maps to Kubernetes ResourceRequirements.
+	// Maximum length for this list is 10.
+	// Minimum length for this list is 1.
+	// Each resource name must be unique within this list.
+	// +optional
+	// +listType=map
+	// +listMapKey=name
+	// +kubebuilder:validation:MaxItems=10
+	// +kubebuilder:validation:MinItems=1
+	Resources []ContainerResource `json:"resources,omitempty"`
+	// topologySpreadConstraints defines rules for how admission webhook Pods should be distributed
+	// across topology domains such as zones, nodes, or other user-defined labels.
+	// topologySpreadConstraints is optional.
+	// This helps improve high availability and resource efficiency by avoiding placing
+	// too many replicas in the same failure domain.
+	//
+	// When omitted, this means no opinion and the platform is left to choose a default, which is subject to change over time.
+	// This field maps directly to the `topologySpreadConstraints` field in the Pod spec.
+	// Default is empty list.
+	// Maximum length for this list is 10.
+	// Minimum length for this list is 1.
+	// Entries must have unique topologyKey and whenUnsatisfiable pairs.
+	// +kubebuilder:validation:MaxItems=10
+	// +kubebuilder:validation:MinItems=1
+	// +listType=map
+	// +listMapKey=topologyKey
+	// +listMapKey=whenUnsatisfiable
+	// +optional
+	TopologySpreadConstraints []v1.TopologySpreadConstraint `json:"topologySpreadConstraints,omitempty"`
+}

Evidence

PrometheusOperatorConfig and MetricsServerConfig both define NodeSelector and Tolerations fields for
pod scheduling control, but the new PrometheusOperatorAdmissionWebhookConfig only exposes Resources
and TopologySpreadConstraints. The admission webhook is a critical component — if nodes require
taints that the webhook pod cannot tolerate, the pod becomes unschedulable.

config/v1alpha1/types_cluster_monitoring.go[437-516]
config/v1alpha1/types_cluster_monitoring.go[518-559]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution
## Issue description
The `PrometheusOperatorAdmissionWebhookConfig` struct is missing `nodeSelector` and `tolerations` fields that exist in the sibling config structs `PrometheusOperatorConfig` and `MetricsServerConfig`. Without these fields, users cannot control which nodes the admission webhook pod is scheduled on, and cannot add tolerations to allow scheduling on tainted nodes. Since the admission webhook validates all `PrometheusRule` and `AlertmanagerConfig` objects, an unschedulable webhook pod would block those operations cluster-wide.
## Issue Context
- `PrometheusOperatorConfig` (lines 437–516) defines `NodeSelector map[string]string` and `Tolerations []v1.Toleration`
- `MetricsServerConfig` (lines 349–432) defines the same fields
- `PrometheusOperatorAdmissionWebhookConfig` (lines 518–559) only defines `Resources` and `TopologySpreadConstraints`
- The CRD and OpenAPI generated files would also need to be regenerated
## Fix Focus Areas
- config/v1alpha1/types_cluster_monitoring.go[518-559]
- config/v1alpha1/zz_generated.deepcopy.go[1066-1094]
- config/v1alpha1/zz_generated.swagger_doc_generated.go[222-230]
- config/v1alpha1/zz_generated.crd-manifests/0000_10_config-operator_01_clustermonitorings-CustomNoUpgrade.crd.yaml[1289-1566]
- config/v1alpha1/zz_generated.crd-manifests/0000_10_config-operator_01_clustermonitorings-DevPreviewNoUpgrade.crd.yaml[1289-1566]
- config/v1alpha1/zz_generated.crd-manifests/0000_10_config-operator_01_clustermonitorings-TechPreviewNoUpgrade.crd.yaml[1289-1566]
- config/v1alpha1/zz_generated.featuregated-crd-manifests/clustermonitorings.config.openshift.io/ClusterMonitoringConfig.yaml[1289-1566]
- openapi/generated_openapi/zz_generated.openapi.go[24099-24157]
- config/v1alpha1/tests/clustermonitorings.config.openshift.io/ClusterMonitoringConfig.yaml[354-553]

i Copy this prompt and use it to remediate the issue with your preferred AI generation tools

Description

The new tests for prometheusOperatorAdmissionWebhookConfig.resources do not include a test case
for duplicate resource names, even though an equivalent test exists for
alertmanagerConfig.customConfig.resources and the CRD enforces uniqueness via
x-kubernetes-list-type: map. This leaves the uniqueness enforcement unvalidated for the new field.

Code

config/v1alpha1/tests/clustermonitorings.config.openshift.io/ClusterMonitoringConfig.yaml[R456-553]

+ - name: Should reject PrometheusOperatorAdmissionWebhookConfig with too many resources
+ initial: |
+ apiVersion: config.openshift.io/v1alpha1
+ kind: ClusterMonitoring
+ spec:
+ prometheusOperatorAdmissionWebhookConfig:
+ resources:
+ - name: "cpu"
+ request: "100m"
+ - name: "memory"
+ request: "64Mi"
+ - name: "hugepages-2Mi"
+ request: "32Mi"
+ - name: "hugepages-1Gi"
+ request: "1Gi"
+ - name: "ephemeral-storage"
+ request: "1Gi"
+ - name: "nvidia.com/gpu"
+ request: "1"
+ - name: "example.com/foo"
+ request: "1"
+ - name: "example.com/bar"
+ request: "1"
+ - name: "example.com/baz"
+ request: "1"
+ - name: "example.com/qux"
+ request: "1"
+ - name: "example.com/quux"
+ request: "1"
+ expectedError: 'spec.prometheusOperatorAdmissionWebhookConfig.resources: Too many: 11: must have at most 10 items'
+ - name: Should reject PrometheusOperatorAdmissionWebhookConfig with limit less than request
+ initial: |
+ apiVersion: config.openshift.io/v1alpha1
+ kind: ClusterMonitoring
+ spec:
+ prometheusOperatorAdmissionWebhookConfig:
+ resources:
+ - name: "cpu"
+ request: "500m"
+ limit: "200m"
+ expectedError: 'spec.prometheusOperatorAdmissionWebhookConfig.resources[0]: Invalid value: "object": limit must be greater than or equal to request'
+ - name: Should reject PrometheusOperatorAdmissionWebhookConfig with too many topologySpreadConstraints
+ initial: |
+ apiVersion: config.openshift.io/v1alpha1
+ kind: ClusterMonitoring
+ spec:
+ prometheusOperatorAdmissionWebhookConfig:
+ topologySpreadConstraints:
+ - maxSkew: 1
+ topologyKey: "zone1"
+ whenUnsatisfiable: DoNotSchedule
+ - maxSkew: 1
+ topologyKey: "zone2"
+ whenUnsatisfiable: DoNotSchedule
+ - maxSkew: 1
+ topologyKey: "zone3"
+ whenUnsatisfiable: DoNotSchedule
+ - maxSkew: 1
+ topologyKey: "zone4"
+ whenUnsatisfiable: DoNotSchedule
+ - maxSkew: 1
+ topologyKey: "zone5"
+ whenUnsatisfiable: DoNotSchedule
+ - maxSkew: 1
+ topologyKey: "zone6"
+ whenUnsatisfiable: DoNotSchedule
+ - maxSkew: 1
+ topologyKey: "zone7"
+ whenUnsatisfiable: DoNotSchedule
+ - maxSkew: 1
+ topologyKey: "zone8"
+ whenUnsatisfiable: DoNotSchedule
+ - maxSkew: 1
+ topologyKey: "zone9"
+ whenUnsatisfiable: DoNotSchedule
+ - maxSkew: 1
+ topologyKey: "zone10"
+ whenUnsatisfiable: DoNotSchedule
+ - maxSkew: 1
+ topologyKey: "zone11"
+ whenUnsatisfiable: DoNotSchedule
+ expectedError: 'spec.prometheusOperatorAdmissionWebhookConfig.topologySpreadConstraints: Too many: 11: must have at most 10 items'
+ - name: Should reject PrometheusOperatorAdmissionWebhookConfig with empty topologySpreadConstraints array
+ initial: |
+ apiVersion: config.openshift.io/v1alpha1
+ kind: ClusterMonitoring
+ spec:
+ prometheusOperatorAdmissionWebhookConfig:
+ topologySpreadConstraints: []
+ expectedError: 'spec.prometheusOperatorAdmissionWebhookConfig.topologySpreadConstraints: Invalid value: 0: spec.prometheusOperatorAdmissionWebhookConfig.topologySpreadConstraints in body should have at least 1 items'
+ - name: Should reject PrometheusOperatorAdmissionWebhookConfig with empty resources array
+ initial: |
+ apiVersion: config.openshift.io/v1alpha1
+ kind: ClusterMonitoring
+ spec:
+ prometheusOperatorAdmissionWebhookConfig:
+ resources: []
+ expectedError: 'spec.prometheusOperatorAdmissionWebhookConfig.resources: Invalid value: 0: spec.prometheusOperatorAdmissionWebhookConfig.resources in body should have at least 1 items'

Evidence

The test file has a 'Should reject ContainerResource with duplicate names' test for
alertmanagerConfig.customConfig.resources (lines 21-36), but the new test block for
prometheusOperatorAdmissionWebhookConfig (lines 354-553 in the diff) covers only: valid resources,
valid topologySpreadConstraints, both fields, empty object, too many resources, limit < request, too
many topologySpreadConstraints, and empty arrays — but not duplicate resource names. The CRD sets
x-kubernetes-list-map-keys: [name] and x-kubernetes-list-type: map for this field, so the validation
exists but is untested.

config/v1alpha1/tests/clustermonitorings.config.openshift.io/ClusterMonitoringConfig.yaml[21-36]
config/v1alpha1/zz_generated.crd-manifests/0000_10_config-operator_01_clustermonitorings-CustomNoUpgrade.crd.yaml[1371-1373]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution
## Issue description
The test suite for `prometheusOperatorAdmissionWebhookConfig` is missing a test case that verifies duplicate resource names are rejected. The `resources` field uses `x-kubernetes-list-type: map` with `x-kubernetes-list-map-keys: [name]`, which enforces uniqueness at the API server level, but this behavior is not explicitly tested for the new field.
## Issue Context
- An equivalent test exists for `alertmanagerConfig.customConfig.resources` at lines 21-36 of the test YAML
- The new tests (lines 354-553 in the diff) cover many error cases but omit the duplicate-name scenario
- The CRD enforces uniqueness via `x-kubernetes-list-map-keys: [name]` and `x-kubernetes-list-type: map`
## Fix Focus Areas
- config/v1alpha1/tests/clustermonitorings.config.openshift.io/ClusterMonitoringConfig.yaml[546-553]
Add the following test case after the existing rejection tests:
```yaml
- name: Should reject PrometheusOperatorAdmissionWebhookConfig with duplicate resource names
 initial: |
 apiVersion: config.openshift.io/v1alpha1
 kind: ClusterMonitoring
 spec:
 prometheusOperatorAdmissionWebhookConfig:
 resources:
 - name: "cpu"
 request: "100m"
 - name: "cpu"
 request: "200m"
 expectedError: 'spec.prometheusOperatorAdmissionWebhookConfig.resources[1]: Duplicate value: map[string]interface {}{"name":"cpu"}'
```

i Copy this prompt and use it to remediate the issue with your preferred AI generation tools

In response to this:

Add configuration options for the Prometheus Operator admission webhook
component in config/v1alpha1.

The new PrometheusOperatorAdmissionWebhookConfig struct supports:

• resources: compute resource requests and limits
• topologySpreadConstraints: pod distribution across topology domains

Signed-off-by: Daniel Mellado dmellado@fedoraproject.org

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the openshift-eng/jira-lifecycle-plugin repository.

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.