--- apiVersion: v1 items: - apiVersion: v1 data: _example: |- ################################ # # # EXAMPLE CONFIGURATION # # # ################################ # This block is not actually functional configuration, # but serves to illustrate the available configuration # options and document them in a way that is accessible # to users that `kubectl edit` this config map. # # These sample configuration options may be copied out of # this example block and unindented to be in the data block # to actually change the configuration. # ====================================== EXPLAINERS CONFIGURATION ====================================== # Example explainers: |- { "art": { "image" : "kserve/art-explainer", "defaultImageVersion": "latest" } } # Art Explainer runtime configuration explainers: |- { # Art explainer runtime configuration "art": { # image contains the default Art explainer serving runtime image uri. "image" : "kserve/art-explainer", # defautltImageVersion contains the Art explainer serving runtime default image version. "defaultImageVersion": "latest" } } # ====================================== ISVC CONFIGURATION ====================================== # Example - setting custom annotation inferenceService: |- { "serviceAnnotationDisallowedList": [ "my.custom.annotation/1" ], "serviceLabelDisallowedList": [ "my.custom.label.1" ] } # Example - setting custom annotation inferenceService: |- { # ServiceAnnotationDisallowedList is a list of annotations that are not allowed to be propagated to Knative # revisions, which prevents the reconciliation loop to be triggered if the annotations is # configured here are used. # Default values are: # "autoscaling.knative.dev/min-scale", # "autoscaling.knative.dev/max-scale", # "internal.serving.kserve.io/storage-initializer-sourceuri", # "kubectl.kubernetes.io/last-applied-configuration", # "modelFormat" # Any new value will be appended to the list. "serviceAnnotationDisallowedList": [ "my.custom.annotation/1" ], # ServiceLabelDisallowedList is a list of labels that are not allowed to be propagated to Knative revisions # which prevents the reconciliation loop to be triggered if the labels is configured here are used. "serviceLabelDisallowedList": [ "my.custom.label.1" ] } # Example - setting custom resource inferenceService: |- { "resource": { "cpuLimit": "1", "memoryLimit": "2Gi", "cpuRequest": "1", "memoryRequest": "2Gi" } } # Example - setting custom resource inferenceService: |- { # resource contains the default resource configuration for the inference service. # you can override this configuration by specifying the resources in the inference service yaml. # If you want to unbound the resource (limits and requests), you can set the value to null or "" # or just remove the specific field from the config. "resource": { # cpuLimit is the limits.cpu to set for the inference service. "cpuLimit": "1", # memoryLimit is the limits.memory to set for the inference service. "memoryLimit": "2Gi", # cpuRequest is the requests.cpu to set for the inference service. "cpuRequest": "1", # memoryRequest is the requests.memory to set for the inference service. "memoryRequest": "2Gi" } } # ====================================== MultiNode CONFIGURATION ====================================== # Example multiNode: |- { "customGPUResourceTypeList": [ "custom.com/gpu" ] } # Example of multinode configuration multiNode: |- { # CustomGPUResourceTypeList is a list of custom GPU resource types intended to identify the GPU type of a resource, # not to restrict the user from using a specific GPU type. # The MultiNode runtime pod will dynamically add GPU resources based on the registered GPU types. "customGPUResourceTypeList": [ "custom.com/gpu" ] } # ====================================== OTelCollector CONFIGURATION ====================================== # Example opentelemetryCollector: |- { # scrapeInterval is the interval at which the OpenTelemetry Collector will scrape the metrics. "scrapeInterval": "5s", # metricScalerEndpoint is the endpoint from which the KEDA's ScaledObject will scrape the metrics. "metricScalerEndpoint": "keda-otel-scaler.keda.svc:4318", # metricReceiverEndpoint is the endpoint from which the OpenTelemetry Collector will scrape the metrics. "metricReceiverEndpoint": "keda-otel-scaler.keda.svc:4317" } # ====================================== AUTOSCALER CONFIGURATION ====================================== # Example autoscaler: |- { # scaleUpStabilizationWindowSeconds is the stabilization window in seconds for scale up. "scaleUpStabilizationWindowSeconds": "0", # scaleDownStabilizationWindowSeconds is the stabilization window in seconds for scale down. "scaleDownStabilizationWindowSeconds": "300" } # ====================================== STORAGE INITIALIZER CONFIGURATION ====================================== # Example storageInitializer: |- { "image" : "kserve/storage-initializer:latest", "memoryRequest": "100Mi", "memoryLimit": "1Gi", "cpuRequest": "100m", "cpuLimit": "1", "caBundleConfigMapName": "", "caBundleVolumeMountPath": "/etc/ssl/custom-certs", "enableModelcar": false, "cpuModelcar": "10m", "memoryModelcar": "15Mi" } storageInitializer: |- { # image contains the default storage initializer image uri. "image" : "kserve/storage-initializer:latest", # memoryRequest is the requests.memory to set for the storage initializer init container. "memoryRequest": "100Mi", # memoryLimit is the limits.memory to set for the storage initializer init container. "memoryLimit": "1Gi", # cpuRequest is the requests.cpu to set for the storage initializer init container. "cpuRequest": "100m", # cpuLimit is the limits.cpu to set for the storage initializer init container. "cpuLimit": "1", # caBundleConfigMapName is the ConfigMap will be copied to a user namespace for the storage initializer init container. "caBundleConfigMapName": "", # caBundleVolumeMountPath is the mount point for the configmap set by caBundleConfigMapName for the storage initializer init container. "caBundleVolumeMountPath": "/etc/ssl/custom-certs", # enableModelcar enabled allows you to directly access an OCI container image by # using a source URL with an "oci://" schema. "enableModelcar": false, # cpuModelcar is the cpu request and limit that is used for the passive modelcar container. It can be # set very low, but should be allowed by any Kubernetes LimitRange that might apply. "cpuModelcar": "10m", # cpuModelcar is the memory request and limit that is used for the passive modelcar container. It can be # set very low, but should be allowed by any Kubernetes LimitRange that might apply. "memoryModelcar": "15Mi", # uidModelcar is the UID under with which the modelcar process and the main container is running. # Some Kubernetes clusters might require this to be root (0). If not set the user id is left untouched (default) "uidModelcar": 10 } # ====================================== CREDENTIALS ====================================== # Example credentials: |- { "storageSpecSecretName": "storage-config", "storageSecretNameAnnotation": "serving.kserve.io/storageSecretName", "gcs": { "gcsCredentialFileName": "gcloud-application-credentials.json" }, "s3": { "s3AccessKeyIDName": "AWS_ACCESS_KEY_ID", "s3SecretAccessKeyName": "AWS_SECRET_ACCESS_KEY", "s3Endpoint": "", "s3UseHttps": "", "s3Region": "", "s3VerifySSL": "", "s3UseVirtualBucket": "", "s3UseAccelerate": "", "s3UseAnonymousCredential": "", "s3CABundleConfigMap": "", "s3CABundle": "" } } # This is a global configuration used for downloading models from the cloud storage. # You can override this configuration by specifying the annotations on service account or static secret. # https://kserve.github.io/website/master/modelserving/storage/s3/s3/ # For a quick reference about AWS ENV variables: # AWS Cli: https://docs.aws.amazon.com/cli/latest/userguide/cli-configure-envvars.html # Boto: https://boto3.amazonaws.com/v1/documentation/api/latest/guide/configuration.html#using-environment-variables # # The `s3AccessKeyIDName` and `s3SecretAccessKeyName` fields are only used from this configmap when static credentials (IAM User Access Key Secret) # are used as the authentication method for AWS S3. # The rest of the fields are used in both authentication methods (IAM Role for Service Account & IAM User Access Key Secret) if a non-empty value is provided. credentials: |- { # storageSpecSecretName contains the secret name which has the credentials for downloading the model. # This option is used when specifying the storage spec on isvc yaml. "storageSpecSecretName": "storage-config", # The annotation can be specified on isvc yaml to allow overriding with the secret name reference from the annotation value. # When using storageUri the order of the precedence is: secret name reference annotation > secret name references from service account # When using storageSpec the order of the precedence is: secret name reference annotation > storageSpecSecretName in configmap # Configuration for google cloud storage "gcs": { # gcsCredentialFileName specifies the filename of the gcs credential "gcsCredentialFileName": "gcloud-application-credentials.json" }, # Configuration for aws s3 storage. This add the corresponding environmental variables to the storage initializer init container. # For more info on s3 storage see https://kserve.github.io/website/master/modelserving/storage/s3/s3/ "s3": { # s3AccessKeyIDName specifies the s3 access key id name "s3AccessKeyIDName": "AWS_ACCESS_KEY_ID", # s3SecretAccessKeyName specifies the s3 secret access key name "s3SecretAccessKeyName": "AWS_SECRET_ACCESS_KEY", # s3Endpoint specifies the s3 endpoint "s3Endpoint": "", # s3UseHttps controls whether to use secure https or unsecure http to download models. # Allowed values are 0 and 1. "s3UseHttps": "", # s3Region specifies the region of the bucket. "s3Region": "", # s3VerifySSL controls whether to verify the tls/ssl certificate. "s3VerifySSL": "", # s3UseVirtualBucket configures whether it is a virtual bucket or not. "s3UseVirtualBucket": "", # s3UseAccelerate configures whether to use transfer acceleration. "s3UseAccelerate": "", # s3UseAnonymousCredential configures whether to use anonymous credentials to download the model or not. "s3UseAnonymousCredential": "", # s3CABundleConfigMap specifies the mounted CA bundle config map name. "s3CABundleConfigMap": "", # s3CABundle specifies the full path (mount path + file name) for the mounted config map data when used with a configured CA bundle config map. # s3CABundle specifies the path to a certificate bundle to use for HTTPS certificate validation when used absent of a configured CA bundle config map. "s3CABundle": "" } } # ====================================== INGRESS CONFIGURATION ====================================== # Example ingress: |- { "enableGatewayApi": false, "kserveIngressGateway": "kserve/kserve-ingress-gateway", "ingressGateway" : "knative-serving/knative-ingress-gateway", "localGateway" : "knative-serving/knative-local-gateway", "localGatewayService" : "knative-local-gateway.istio-system.svc.cluster.local", "ingressDomain" : "example.com", "additionalIngressDomains": ["additional-example.com", "additional-example-1.com"], "ingressClassName" : "istio", "domainTemplate": "{{ .Name }}-{{ .Namespace }}.{{ .IngressDomain }}", "urlScheme": "http", "disableIstioVirtualHost": false, "disableIngressCreation": false, "disableHTTPRouteTimeout": false } ingress: |- { # enableGatewayApi specifies whether to use Gateway API instead of Ingress to serve external traffic. "enableGatewayApi": false, # KServe implements [Gateway API](https://gateway-api.sigs.k8s.io/) to serve external traffic. # By default, KServe configures a default gateway to serve external traffic. # But, KServe can be configured to use a custom gateway by modifying this configuration. # The gateway should be specified in format / # NOTE: This configuration only applicable for raw deployment. "kserveIngressGateway": "kserve/kserve-ingress-gateway", # ingressGateway specifies the ingress gateway to serve external traffic. # The gateway should be specified in format / # NOTE: This configuration only applicable for serverless deployment with Istio configured as network layer. "ingressGateway" : "knative-serving/knative-ingress-gateway", # knativeLocalGatewayService specifies the hostname of the Knative's local gateway service. # The default KServe configurations are re-using the Istio local gateways for Knative. In this case, this # knativeLocalGatewayService field can be left unset. When unset, the value of "localGatewayService" will be used. # However, sometimes it may be better to have local gateways specifically for KServe (e.g. when enabling strict mTLS in Istio). # Under such setups where KServe is needed to have its own local gateways, the values of the "localGateway" and # "localGatewayService" should point to the KServe local gateways. Then, this knativeLocalGatewayService field # should point to the Knative's local gateway service. # NOTE: This configuration only applicable for serverless deployment with Istio configured as network layer. "knativeLocalGatewayService": "", # localGateway specifies the gateway which handles the network traffic within the cluster. # NOTE: This configuration only applicable for serverless deployment with Istio configured as network layer. "localGateway" : "knative-serving/knative-local-gateway", # localGatewayService specifies the hostname of the local gateway service. # NOTE: This configuration only applicable for serverless deployment with Istio configured as network layer. "localGatewayService" : "knative-local-gateway.istio-system.svc.cluster.local", # ingressDomain specifies the domain name which is used for creating the url. # If ingressDomain is empty then example.com is used as default domain. # NOTE: This configuration only applicable for raw deployment. "ingressDomain" : "example.com", # additionalIngressDomains specifies the additional domain names which are used for creating the url. "additionalIngressDomains": ["additional-example.com", "additional-example-1.com"] # ingressClassName specifies the ingress controller to use for ingress traffic. # This is optional and if omitted the default ingress in the cluster is used. # https://kubernetes.io/docs/concepts/services-networking/ingress/#default-ingress-class # NOTE: This configuration only applicable for raw deployment. "ingressClassName" : "istio", # domainTemplate specifies the template for generating domain/url for each inference service by combining variable from: # Name of the inference service ( {{ .Name}} ) # Namespace of the inference service ( {{ .Namespace }} ) # Annotation of the inference service ( {{ .Annotations.key }} ) # Label of the inference service ( {{ .Labels.key }} ) # IngressDomain ( {{ .IngressDomain }} ) # If domain template is empty the default template {{ .Name }}-{{ .Namespace }}.{{ .IngressDomain }} is used. # NOTE: This configuration only applicable for raw deployment. "domainTemplate": "{{ .Name }}-{{ .Namespace }}.{{ .IngressDomain }}", # urlScheme specifies the url scheme to use for inference service and inference graph. # If urlScheme is empty then by default http is used. "urlScheme": "http", # disableIstioVirtualHost controls whether to use istio as network layer. # By default istio is used as the network layer. When DisableIstioVirtualHost is true, KServe does not # create the top level virtual service thus Istio is no longer required for serverless mode. # By setting this field to true, user can use other networking layers supported by knative. # For more info https://github.com/kserve/kserve/pull/2380, https://kserve.github.io/website/master/admin/serverless/kourier_networking/. # NOTE: This configuration is only applicable to serverless deployment. "disableIstioVirtualHost": false, # disableIngressCreation controls whether to disable ingress creation for raw deployment mode. "disableIngressCreation": false, # disableHTTPRouteTimeout controls whether to omit the timeout field from HTTPRoute rules. # Set to true for Gateway controllers (e.g. GKE Gateway) that do not support the optional timeouts field. "disableHTTPRouteTimeout": false, # pathTemplate specifies the template for generating path based url for each inference service. # The following variables can be used in the template for generating url. # Name of the inference service ( {{ .Name}} ) # Namespace of the inference service ( {{ .Namespace }} ) # For more info https://github.com/kserve/kserve/issues/2257. # NOTE: This configuration only applicable to serverless deployment. "pathTemplate": "/serving/{{ .Namespace }}/{{ .Name }}" } # ====================================== LOGGER CONFIGURATION ====================================== # Example logger: |- { "image" : "kserve/agent:latest", "memoryRequest": "100Mi", "memoryLimit": "1Gi", "cpuRequest": "100m", "cpuLimit": "1", "defaultUrl": "http://default-broker" } logger: |- { # image contains the default logger image uri. "image" : "kserve/agent:latest", # memoryRequest is the requests.memory to set for the logger container. "memoryRequest": "100Mi", # memoryLimit is the limits.memory to set for the logger container. "memoryLimit": "1Gi", # cpuRequest is the requests.cpu to set for the logger container. "cpuRequest": "100m", # cpuLimit is the limits.cpu to set for the logger container. "cpuLimit": "1", # defaultUrl specifies the default logger url. If logger is not specified in the resource this url is used. "defaultUrl": "http://default-broker" } # ====================================== BATCHER CONFIGURATION ====================================== # Example batcher: |- { "image" : "kserve/agent:latest", "memoryRequest": "1Gi", "memoryLimit": "1Gi", "cpuRequest": "1", "cpuLimit": "1", "maxBatchSize": "32", "maxLatency": "5000" } batcher: |- { # image contains the default batcher image uri. "image" : "kserve/agent:latest", # memoryRequest is the requests.memory to set for the batcher container. "memoryRequest": "1Gi", # memoryLimit is the limits.memory to set for the batcher container. "memoryLimit": "1Gi", # cpuRequest is the requests.cpu to set for the batcher container. "cpuRequest": "1", # cpuLimit is the limits.cpu to set for the batcher container. "cpuLimit": "1" # maxBatchSize is the default maximum batch size for batcher. "maxBatchSize": "32", # maxLatency is the default maximum latency in milliseconds for batcher to wait and collect the batch. "maxLatency": "5000" } # ====================================== AGENT CONFIGURATION ====================================== # Example agent: |- { "image" : "kserve/agent:latest", "memoryRequest": "100Mi", "memoryLimit": "1Gi", "cpuRequest": "100m", "cpuLimit": "1" } agent: |- { # image contains the default agent image uri. "image" : "kserve/agent:latest", # memoryRequest is the requests.memory to set for the agent container. "memoryRequest": "100Mi", # memoryLimit is the limits.memory to set for the agent container. "memoryLimit": "1Gi", # cpuRequest is the requests.cpu to set for the agent container. "cpuRequest": "100m", # cpuLimit is the limits.cpu to set for the agent container. "cpuLimit": "1" } # ====================================== ROUTER CONFIGURATION ====================================== # Example router: |- { "image" : "kserve/router:latest", "memoryRequest": "100Mi", "memoryLimit": "1Gi", "cpuRequest": "100m", "cpuLimit": "1", "headers": { "propagate": [] }, "imagePullPolicy": "IfNotPresent", "imagePullSecrets": ["docker-secret"] } # router is the implementation of inference graph. router: |- { # image contains the default router image uri. "image" : "kserve/router:latest", # memoryRequest is the requests.memory to set for the router container. "memoryRequest": "100Mi", # memoryLimit is the limits.memory to set for the router container. "memoryLimit": "1Gi", # cpuRequest is the requests.cpu to set for the router container. "cpuRequest": "100m", # cpuLimit is the limits.cpu to set for the router container. "cpuLimit": "1", # Propagate the specified headers to all the steps specified in an InferenceGraph. # You can either specify the exact header names or use [Golang supported regex patterns] # (https://pkg.go.dev/regexp/syntax@go1.21.3#hdr-Syntax) to propagate multiple headers. "headers": { "propagate": [ "Authorization", "Test-Header-*", "*Trace-Id*" ] } # imagePullPolicy specifies when the router image should be pulled from registry. "imagePullPolicy": "IfNotPresent", # # imagePullSecrets specifies the list of secrets to be used for pulling the router image from registry. # https://kubernetes.io/docs/tasks/configure-pod-container/pull-image-private-registry/ "imagePullSecrets": ["docker-secret"] } # ====================================== DEPLOYMENT CONFIGURATION ====================================== # Example deploy: |- { "defaultDeploymentMode": "Serverless", "deploymentRolloutStrategy": { "defaultRollout": { "maxSurge": "1", "maxUnavailable": "1" } } } deploy: |- { # defaultDeploymentMode specifies the default deployment mode of the kserve. The supported values are # Standard and Knative. Users can override the deployment mode at service level # by adding the annotation serving.kserve.io/deploymentMode. # "defaultDeploymentMode": "Standard", # deploymentRolloutStrategy specifies the default rollout strategy for the Standard deployment mode # "deploymentRolloutStrategy": { # defaultRollout specifies the default rollout configuration using Kubernetes deployment strategy # "defaultRollout": { # maxSurge specifies the maximum number of pods that can be created above the desired replica count # Can be an absolute number (ex: 5) or a percentage of desired pods (ex: 10%) # "maxSurge": "1", # maxUnavailable specifies the maximum number of pods that can be unavailable during the update # Can be an absolute number (ex: 5) or a percentage of desired pods (ex: 10%) # "maxUnavailable": "1" # } # } } # ====================================== SERVICE CONFIGURATION ====================================== # Example service: |- { "serviceClusterIPNone": false } service: |- { # ServiceClusterIPNone is a boolean flag to indicate if the service should have a clusterIP set to None. # If the DeploymentMode is Raw, the default value for ServiceClusterIPNone if not set is false # "serviceClusterIPNone": false } # ====================================== METRICS CONFIGURATION ====================================== # Example metricsAggregator: |- { "enableMetricAggregation": "false", "enablePrometheusScraping" : "false" } # For more info see https://github.com/kserve/kserve/blob/master/qpext/README.md metricsAggregator: |- { # enableMetricAggregation configures metric aggregation annotation. This adds the annotation serving.kserve.io/enable-metric-aggregation to every # service with the specified boolean value. If true enables metric aggregation in queue-proxy by setting env vars in the queue proxy container # to configure scraping ports. "enableMetricAggregation": "false", # enablePrometheusScraping configures metric aggregation annotation. This adds the annotation serving.kserve.io/enable-metric-aggregation to every # service with the specified boolean value. If true, prometheus annotations are added to the pod. If serving.kserve.io/enable-metric-aggregation is false, # the prometheus port is set with the default prometheus scraping port 9090, otherwise the prometheus port annotation is set with the metric aggregation port. "enablePrometheusScraping" : "false" } # ====================================== LOCALMODEL CONFIGURATION ====================================== # Example localModel: |- { "enabled": false, # jobNamespace specifies the namespace where the download job will be created. "jobNamespace": "kserve-localmodel-jobs", # defaultJobImage specifies the default image used for the download job. "defaultJobImage" : "kserve/storage-initializer:latest", # Kubernetes modifies the filesystem group ID on the attached volume. "fsGroup": 1000, # TTL for the download job after it is finished. "jobTTLSecondsAfterFinished": 3600, # The frequency at which the local model agent reconciles the local models # This is to detect if models are missing from local disk "reconcilationFrequencyInSecs": 60, # This is to disable localmodel pv and pvc management for namespaces without isvcs "disableVolumeManagement": false } agent: |- { "cpuLimit": "1", "cpuRequest": "100m", "image": "quay.io/opendatahub/kserve-agent@sha256:41e9c610ce929fc56f4076f97bf42d251460cf8b6a90f49e0a84e5b92946522f", "memoryLimit": "1Gi", "memoryRequest": "100Mi" } autoscaler: |- { "scaleUpStabilizationWindowSeconds": "0", "scaleDownStabilizationWindowSeconds": "300" } autoscaling-wva-controller-config: |- { "prometheus": { "url": "https://thanos-querier.openshift-monitoring.svc.cluster.local:9091", "authModes": "bearer", "triggerAuthName": "ai-inference-keda-thanos", "triggerAuthKind": "ClusterTriggerAuthentication" } } batcher: |- { "cpuLimit": "1", "cpuRequest": "1", "image": "quay.io/opendatahub/kserve-agent@sha256:41e9c610ce929fc56f4076f97bf42d251460cf8b6a90f49e0a84e5b92946522f", "memoryLimit": "1Gi", "memoryRequest": "1Gi" } credentials: |- { "storageSpecSecretName": "storage-config", "storageSecretNameAnnotation": "serving.kserve.io/storageSecretName", "gcs": { "gcsCredentialFileName": "gcloud-application-credentials.json" }, "s3": { "s3AccessKeyIDName": "AWS_ACCESS_KEY_ID", "s3SecretAccessKeyName": "AWS_SECRET_ACCESS_KEY", "s3Endpoint": "", "s3UseHttps": "", "s3Region": "", "s3VerifySSL": "", "s3UseVirtualBucket": "", "s3UseAccelerate": "", "s3UseAnonymousCredential": "", "s3CABundleConfigMap": "odh-kserve-custom-ca-bundle", "s3CABundle": "/etc/ssl/custom-certs/cabundle.crt" } } deploy: |- { "defaultDeploymentMode": "RawDeployment" } explainers: '{}' inferenceService: |- { "serviceAnnotationDisallowedList": [ "autoscaling.knative.dev/min-scale", "autoscaling.knative.dev/max-scale", "internal.serving.kserve.io/storage-initializer-sourceuri", "kubectl.kubernetes.io/last-applied-configuration", "security.opendatahub.io/enable-auth", "networking.knative.dev/visibility", "haproxy.router.openshift.io/timeout", "opendatahub.io/hardware-profile-name", "opendatahub.io/hardware-profile-namespace" ] } ingress: |- { "enableGatewayApi": false, "kserveIngressGateway": "openshift-ingress/openshift-ai-inference", "enableLLMInferenceServiceTLS": true, "ingressGateway": "knative-serving/knative-ingress-gateway", "knativeLocalGatewayService": "knative-local-gateway.istio-system.svc.cluster.local", "ingressService": "istio-ingressgateway.istio-system.svc.cluster.local", "localGateway": "istio-system/kserve-local-gateway", "localGatewayService": "kserve-local-gateway.istio-system.svc.cluster.local", "ingressDomain": "apps.edb64d1a-b546-40c9-aa62-6406bece498e.prod.konfluxeaas.com", "ingressClassName": "openshift-default", "domainTemplate": "{{ .Name }}-{{ .Namespace }}.{{ .IngressDomain }}", "urlScheme": "http", "disableIstioVirtualHost": false, "disableIngressCreation": true } localModel: |- { "enabled": false, "jobNamespace": "opendatahub", "defaultJobImage" : "REPLACE_IMAGE", "fsGroup": 1000, "localModelAgentImage": "REPLACE_IMAGE", "localModelAgentCpuRequest": "100m", "localModelAgentMemoryRequest": "200Mi", "localModelAgentCpuLimit": "100m", "localModelAgentMemoryLimit": "300Mi" } logger: |- { "cpuLimit": "1", "cpuRequest": "100m", "defaultUrl": "http://default-broker", "image": "quay.io/opendatahub/kserve-agent@sha256:41e9c610ce929fc56f4076f97bf42d251460cf8b6a90f49e0a84e5b92946522f", "memoryLimit": "1Gi", "memoryRequest": "100Mi" } metricsAggregator: |- { "enableMetricAggregation": "false", "enablePrometheusScraping" : "false" } oauthProxy: |- { "cpuLimit": "200m", "cpuRequest": "100m", "image": "quay.io/opendatahub/odh-kube-auth-proxy@sha256:dcb09fbabd8811f0956ef612a0c9ddd5236804b9bd6548a0647d2b531c9d01b3", "memoryLimit": "128Mi", "memoryRequest": "64Mi" } openshiftConfig: |- { "modelcachePermissionFixImage": "REPLACE_IMAGE" } opentelemetryCollector: |- { "scrapeInterval": "5s", "metricReceiverEndpoint": "keda-otel-scaler.keda.svc:4317", "metricScalerEndpoint": "keda-otel-scaler.keda.svc:4318", "resource": { "cpuLimit": "1", "memoryLimit": "2Gi", "cpuRequest": "200m", "memoryRequest": "512Mi" } } router: |- { "cpuLimit": "1", "cpuRequest": "100m", "headers": { "propagate": [ "Authorization" ] }, "image": "quay.io/opendatahub/kserve-router@sha256:c92ab94c818c88d4324be5bb5a3d9bda3e4173010dc47851ff3051c5a06f1614", "memoryLimit": "1Gi", "memoryRequest": "100Mi" } security: |- { "autoMountServiceAccountToken": false } service: |- { "serviceClusterIPNone": false } storageInitializer: |- { "cpuLimit": "1", "cpuModelcar": "10m", "cpuRequest": "100m", "enableModelcar": true, "image": "quay.io/opendatahub/kserve-storage-initializer@sha256:ba8edcbfb3f9312d158be16483785d7654e60c7090f262c42214fd2b29effada", "memoryLimit": "24Gi", "memoryModelcar": "15Mi", "memoryRequest": "100Mi" } kind: ConfigMap metadata: creationTimestamp: "2026-06-15T05:56:30Z" managedFields: - apiVersion: v1 fieldsType: FieldsV1 fieldsV1: f:data: f:_example: {} f:agent: {} f:autoscaler: {} f:autoscaling-wva-controller-config: {} f:batcher: {} f:credentials: {} f:deploy: {} f:explainers: {} f:inferenceService: {} f:localModel: {} f:logger: {} f:metricsAggregator: {} f:oauthProxy: {} f:openshiftConfig: {} f:opentelemetryCollector: {} f:router: {} f:security: {} f:service: {} f:storageInitializer: {} manager: kubectl operation: Apply time: "2026-06-15T05:56:30Z" - apiVersion: v1 fieldsType: FieldsV1 fieldsV1: f:data: f:ingress: {} manager: kubectl-patch operation: Update time: "2026-06-15T05:57:08Z" name: inferenceservice-config namespace: kserve resourceVersion: "13595" uid: c8bf61de-9003-461a-beea-e062dd0748e5 - apiVersion: v1 data: kserve-agent: quay.io/opendatahub/kserve-agent@sha256:41e9c610ce929fc56f4076f97bf42d251460cf8b6a90f49e0a84e5b92946522f kserve-controller: quay.io/opendatahub/kserve-controller@sha256:9fb811904a32cf8c473f3294497ae3de2880b3821ebeb2ef9ae45e81dcbf43d6 kserve-llm-d: registry.redhat.io/rhaiis/vllm-cuda-rhel9@sha256:fc68d623d1bfc36c8cb2fe4a71f19c8578cfb420ce8ce07b20a02c1ee0be0cf3 kserve-llm-d-amd-rocm: registry.redhat.io/rhaiis/vllm-rocm-rhel9@sha256:d9a48add238cc095fa43eeee17c8c4d104de60c4dc623e0bc7f8c4b53b2b2e97 kserve-llm-d-ibm-spyre: registry.redhat.io/rhaiis/vllm-spyre-rhel9@sha256:80ae3e435a5be2c1f117f36599103ab05357917dd6e37f0df6613cb3ac2c13ea kserve-llm-d-inference-scheduler: quay.io/opendatahub/llm-d-inference-scheduler:odh-v3.5-EA1 kserve-llm-d-intel-gaudi: registry.redhat.io/rhaii-early-access/vllm-gaudi-rhel9:3.4.0-ea.2 kserve-llm-d-nvidia-cuda: registry.redhat.io/rhaiis/vllm-cuda-rhel9@sha256:fc68d623d1bfc36c8cb2fe4a71f19c8578cfb420ce8ce07b20a02c1ee0be0cf3 kserve-llm-d-routing-sidecar: quay.io/opendatahub/llm-d-routing-sidecar:odh-stable kserve-llm-d-uds-tokenizer: quay.io/opendatahub/llm-d-kv-cache:v0.7.1 kserve-localmodel-controller: quay.io/opendatahub/odh-kserve-localmodel-controller:odh-stable kserve-localmodelnode-agent: quay.io/opendatahub/odh-kserve-localmodelnode-agent:odh-stable kserve-router: quay.io/opendatahub/kserve-router@sha256:c92ab94c818c88d4324be5bb5a3d9bda3e4173010dc47851ff3051c5a06f1614 kserve-storage-initializer: quay.io/opendatahub/kserve-storage-initializer@sha256:ba8edcbfb3f9312d158be16483785d7654e60c7090f262c42214fd2b29effada kube-rbac-proxy: quay.io/opendatahub/odh-kube-auth-proxy@sha256:dcb09fbabd8811f0956ef612a0c9ddd5236804b9bd6548a0647d2b531c9d01b3 llmisvc-controller: quay.io/opendatahub/odh-kserve-llmisvc-controller:odh-stable kind: ConfigMap metadata: creationTimestamp: "2026-06-15T05:56:30Z" managedFields: - apiVersion: v1 fieldsType: FieldsV1 fieldsV1: f:data: f:kserve-agent: {} f:kserve-controller: {} f:kserve-llm-d: {} f:kserve-llm-d-amd-rocm: {} f:kserve-llm-d-ibm-spyre: {} f:kserve-llm-d-inference-scheduler: {} f:kserve-llm-d-intel-gaudi: {} f:kserve-llm-d-nvidia-cuda: {} f:kserve-llm-d-routing-sidecar: {} f:kserve-llm-d-uds-tokenizer: {} f:kserve-localmodel-controller: {} f:kserve-localmodelnode-agent: {} f:kserve-router: {} f:kserve-storage-initializer: {} f:kube-rbac-proxy: {} f:llmisvc-controller: {} manager: kubectl operation: Apply time: "2026-06-15T05:56:30Z" name: kserve-parameters namespace: kserve resourceVersion: "13192" uid: 9f334bf0-8a28-4d16-afb6-7f42689e9cc3 - apiVersion: v1 data: ca.crt: | -----BEGIN CERTIFICATE----- MIIDPDCCAiSgAwIBAgIIWKqXA278JV0wDQYJKoZIhvcNAQELBQAwJjESMBAGA1UE CxMJb3BlbnNoaWZ0MRAwDgYDVQQDEwdyb290LWNhMB4XDTI2MDYxNTA1MzkxMFoX DTM2MDYxMjA1MzkxMFowJjESMBAGA1UECxMJb3BlbnNoaWZ0MRAwDgYDVQQDEwdy b290LWNhMIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEA0ZggTtu8ZumR FWRqZC6oCmxJrv01EtpqLnWDMwcICakCVmHyAtl868rFbwe51AH5Hd82bYR58OIM tnB/Ql1hw7096sG5YU203yUrtd3t4rJDl7iUzHcfD5AhsX4QJkw5+T7+8yVEx1UD TxBJ+qzygV8o+Go6yk/DH90c08FjZwHJ4Q+5Nwwd3vfhvaW5rmU9+DH6TM3ioMyS 59seVcq9i1rgQcnfD4w+kkH3yPHjkgzfwYqkX8vS+vnygChnDI9VrBX4KOr02UMt FbjFznDpxeCcUANLXXE2kXVQt8s8tBVGmsubzBfQNe4srHV0E3MfArUW5pJwLi1q ShM6JwEZcQIDAQABo24wbDAOBgNVHQ8BAf8EBAMCAqQwDwYDVR0TAQH/BAUwAwEB /zBJBgNVHQ4EQgRA6BEWuWJzT0evgDgD4F8iQ1wUxhpnCd6DUc+RaVwwXi1PUSkD efZgsMVnLSb2G1cFXDhoS2SYOsUS2xv9tBeKLzANBgkqhkiG9w0BAQsFAAOCAQEA aywYfb6/J9Tak8tQ7rVMuyHG0Zi0scZ5jpjaOOgUdpHcquI0+Duts6lAHxj1TmYH +9znORsHh0MgdnK5ng6MKdEY7VzFKEvhdjC5llKCgnhCITVnIgh9JwP+bPCwNhcw 7kYgtcERK4MVkTi1U6wac/ebqEjGGyi4O4w3PJuVLHwyJ1gIIjqgWjH5nxLmw7ci aCW4R09v05xS04OdxcLDhCuP409/l2r2nsjWK7FbwrWYiBwUgIJ3EPa546OK15qq f/WQu6jRFccjCU9bRes7XY+bbcaFUJUPuLuYa4P2b2xH0MpeC99K5oyDvMkSNNzU M6ozEtdsVIU4K0vV0UyIkQ== -----END CERTIFICATE----- -----BEGIN CERTIFICATE----- MIIEADCCAuigAwIBAgIIHl2ozgIAveEwDQYJKoZIhvcNAQELBQAwJjESMBAGA1UE CxMJb3BlbnNoaWZ0MRAwDgYDVQQDEwdyb290LWNhMB4XDTI2MDYxNTA1Mzk0OFoX DTI3MDYxNTA1Mzk0OFowMDESMBAGA1UEChMJb3BlbnNoaWZ0MRowGAYDVQQDExFv cGVuc2hpZnQtaW5ncmVzczCCASIwDQYJKoZIhvcNAQEBBQADggEPADCCAQoCggEB AK4I/ucscvJq55XZaKwLD9bEDET4klS9URuCIe6r53hT4YwSMLUTug22Z1xd6fKw QLuCOka7uhnbfuTKLZeYwqKcdxqtit/YCDwCoIRY+0Ux6d9WFSuLuBr2msCCTdOC ip+/RvzOARV7Lr9+Wfm/cZ26/ADiIntIpNI33qQg9PyMzHajNIftG0ycdVvtuFIX dn3mFHoCKXb7awDNvmXW2jjuulRQPJEAFLo5tXQGHAOt9EcF667Q4WHoJQUBdcHr v/KyUa9R4nKiyK/d39V5XMLiMCCmD3YSWOwbnjPhlgWcBwjOcwBpZ8RL4yKqTqKA CQYAmgyQCgTIVMnUtOr5DfECAwEAAaOCASYwggEiMA4GA1UdDwEB/wQEAwIFoDAd BgNVHSUEFjAUBggrBgEFBQcDAgYIKwYBBQUHAwEwDAYDVR0TAQH/BAIwADBJBgNV HQ4EQgRAeHIlDtjcUy2vIrjar6qJ8Uffk36RzT49jYu8+VpF0Qv2mrnaAEadyOhP AIzo7xwQN56UWesoqnFApbgLDuRtNjBLBgNVHSMERDBCgEDoERa5YnNPR6+AOAPg XyJDXBTGGmcJ3oNRz5FpXDBeLU9RKQN59mCwxWctJvYbVwVcOGhLZJg6xRLbG/20 F4ovMEsGA1UdEQREMEKCQCouYXBwcy5lZGI2NGQxYS1iNTQ2LTQwYzktYWE2Mi02 NDA2YmVjZTQ5OGUucHJvZC5rb25mbHV4ZWFhcy5jb20wDQYJKoZIhvcNAQELBQAD ggEBAD2K2KVd6rTbeCcHLyzgFCncJRAj8yFSIeqBKcEoKjADpUFCqWY2SGg1coxH rsOn0v/qCXf+JmsD/fKMd+QLPtZXx71kzpmwca7ykQRJpVFbqzJLMVBzViBzFMNQ 50GOIrI51WEochNT3wzTa+Da07ZTa7tLOWpjJwG2xPMvuLxb0WAK78yFaiPEkh0O ESqWYfr0Osx5vHmHoCAdQHs9mWTNa5umdaTlJq/zHdMte4jnOdygqTTXb1VFD0E+ LVCmq30PujxBYWmWwv6e7RS20NfRAmvO9wtDDUDEVNYIANoDtKHhxPISVzl6EvjS tUwCURETuWstzWt/ijnMjqFkbvQ= -----END CERTIFICATE----- kind: ConfigMap metadata: annotations: kubernetes.io/description: Contains a CA bundle that can be used to verify the kube-apiserver when using internal endpoints such as the internal service IP or kubernetes.default.svc. No other usage is guaranteed across distributions of Kubernetes clusters. creationTimestamp: "2026-06-15T05:54:52Z" managedFields: - apiVersion: v1 fieldsType: FieldsV1 fieldsV1: f:data: .: {} f:ca.crt: {} f:metadata: f:annotations: .: {} f:kubernetes.io/description: {} manager: kube-controller-manager operation: Update time: "2026-06-15T05:54:52Z" name: kube-root-ca.crt namespace: kserve resourceVersion: "12064" uid: 2801f72d-4666-4b50-b7ca-d51c9d6057b6 - apiVersion: v1 data: cabundle.crt: |- -----BEGIN CERTIFICATE----- MIIDUTCCAjmgAwIBAgIIZETVlxWxI78wDQYJKoZIhvcNAQELBQAwNjE0MDIGA1UE Awwrb3BlbnNoaWZ0LXNlcnZpY2Utc2VydmluZy1zaWduZXJAMTc4MTUwMjczMDAe Fw0yNjA2MTUwNTUyMDlaFw0yODA4MTMwNTUyMTBaMDYxNDAyBgNVBAMMK29wZW5z aGlmdC1zZXJ2aWNlLXNlcnZpbmctc2lnbmVyQDE3ODE1MDI3MzAwggEiMA0GCSqG SIb3DQEBAQUAA4IBDwAwggEKAoIBAQDjTXJFMsjKJTQZCjvkQfgeO/XvPMYsbWnF kIFbVtdeBsMrNQ7g80EAkL2J+rA1lWV5zmCKGVJG+tVHviHj8U7wV2YUo+W/Z+/h zCfwE7CBCQH6RUMhv1bb1KDHrPTHiwrc0Y1pg8scjtuxfhH4c/OHBzTYU8wq+/+A FBsrueltXJwq+sTFk6kfqypWYqj0KxM6T3iWACyAxnFX+5aBbXtjhQNY4SphhQql jQA+dH/dhz/z32AvKDxT5jv9z4mpQziB6LggJW9qWqMYVoRFKKb0FaScftZ7j4lh 4IvUyLqTkhWt4ebS4kkXMdoZFJ7fMtaWjuRKhY+0/MBEt5l4JyPtAgMBAAGjYzBh MA4GA1UdDwEB/wQEAwICpDAPBgNVHRMBAf8EBTADAQH/MB0GA1UdDgQWBBTaOuPH Zy/1pBaOl044n4tzc99fwjAfBgNVHSMEGDAWgBTaOuPHZy/1pBaOl044n4tzc99f wjANBgkqhkiG9w0BAQsFAAOCAQEAnyGfVnvkD5OY3rsba7OASXO4qTzitGuImjQz 9zFa2xLLaNqrXWK/e1Kawi6D/DBuZTcWFwAScWHBV+0SQo9Q35x6mYfPcIzPmJig 9BZLkCN2vStONv2bCr713Izi9vBHSYzbyBI1QZnGz93AhG8YNZUBI06VywGR04Vz 7mAWuy9SoQ+tVe8179n+piKkK/LYChmz2uzjamqjJjW9frslPJz0bAxgQH/RG78J tk59/7Wvtqg/FNsKCPBdTjtNRkQ0niBGXxjp4/dsnTCQMFfN9JCXUj9iwekpz96Q OSpWxEA3lE3QkxjLcPlhbWjw6fwj+Q5GEDFSf0x+87ipy/uYuA== -----END CERTIFICATE----- kind: ConfigMap metadata: creationTimestamp: "2026-06-15T05:57:47Z" labels: opendatahub.io/managed: "true" managedFields: - apiVersion: v1 fieldsType: FieldsV1 fieldsV1: f:data: .: {} f:cabundle.crt: {} f:metadata: f:labels: .: {} f:opendatahub.io/managed: {} manager: manager operation: Update time: "2026-06-15T05:57:47Z" name: odh-kserve-custom-ca-bundle namespace: kserve resourceVersion: "14060" uid: b4ccf5ab-b322-4d90-806d-462a23e81fcf - apiVersion: v1 data: guardrails-detector-huggingface-runtime-image: quay.io/trustyai/guardrails-detector-huggingface-runtime:latest kserve-state: managed mlserver-image: quay.io/opendatahub/mlserver:fast modelregistry-state: removed nim-state: managed odh-model-controller: quay.io/opendatahub/odh-model-controller:fast odh-model-serving-api: quay.io/opendatahub/odh-model-controller:odh-model-serving-api-fast ovms-image: quay.io/opendatahub/openvino_model_server:2025.1-release ray-tls-generator-image: registry.redhat.io/ubi9/ubi-minimal:latest tgis-image: quay.io/opendatahub/text-generation-inference:fast vllm-cpu-image: quay.io/vllm/vllm:latest vllm-cpu-x86-image: quay.io/vllm/vllm:latest vllm-cuda-image: quay.io/vllm/vllm-cuda:latest vllm-gaudi-image: quay.io/opendatahub/vllm:fast-gaudi vllm-rocm-image: quay.io/vllm/vllm-rocm:latest vllm-spyre-image: quay.io/vllm/vllm:latest kind: ConfigMap metadata: annotations: kubectl.kubernetes.io/last-applied-configuration: | {"apiVersion":"v1","data":{"guardrails-detector-huggingface-runtime-image":"quay.io/trustyai/guardrails-detector-huggingface-runtime:latest","kserve-state":"managed","mlserver-image":"quay.io/opendatahub/mlserver:fast","modelregistry-state":"removed","nim-state":"managed","odh-model-controller":"quay.io/opendatahub/odh-model-controller:fast","odh-model-serving-api":"quay.io/opendatahub/odh-model-controller:odh-model-serving-api-fast","ovms-image":"quay.io/opendatahub/openvino_model_server:2025.1-release","ray-tls-generator-image":"registry.redhat.io/ubi9/ubi-minimal:latest","tgis-image":"quay.io/opendatahub/text-generation-inference:fast","vllm-cpu-image":"quay.io/vllm/vllm:latest","vllm-cpu-x86-image":"quay.io/vllm/vllm:latest","vllm-cuda-image":"quay.io/vllm/vllm-cuda:latest","vllm-gaudi-image":"quay.io/opendatahub/vllm:fast-gaudi","vllm-rocm-image":"quay.io/vllm/vllm-rocm:latest","vllm-spyre-image":"quay.io/vllm/vllm:latest"},"kind":"ConfigMap","metadata":{"annotations":{},"name":"odh-model-controller-parameters","namespace":"kserve"}} creationTimestamp: "2026-06-15T05:57:41Z" managedFields: - apiVersion: v1 fieldsType: FieldsV1 fieldsV1: f:data: .: {} f:guardrails-detector-huggingface-runtime-image: {} f:kserve-state: {} f:mlserver-image: {} f:modelregistry-state: {} f:nim-state: {} f:odh-model-controller: {} f:odh-model-serving-api: {} f:ovms-image: {} f:ray-tls-generator-image: {} f:tgis-image: {} f:vllm-cpu-image: {} f:vllm-cpu-x86-image: {} f:vllm-cuda-image: {} f:vllm-gaudi-image: {} f:vllm-rocm-image: {} f:vllm-spyre-image: {} f:metadata: f:annotations: .: {} f:kubectl.kubernetes.io/last-applied-configuration: {} manager: kubectl-client-side-apply operation: Update time: "2026-06-15T05:57:41Z" name: odh-model-controller-parameters namespace: kserve resourceVersion: "13875" uid: 45e06dbf-4d3d-44ee-9fad-3ee1e2aa4764 - apiVersion: v1 data: service-ca.crt: | -----BEGIN CERTIFICATE----- MIIDUTCCAjmgAwIBAgIIZETVlxWxI78wDQYJKoZIhvcNAQELBQAwNjE0MDIGA1UE Awwrb3BlbnNoaWZ0LXNlcnZpY2Utc2VydmluZy1zaWduZXJAMTc4MTUwMjczMDAe Fw0yNjA2MTUwNTUyMDlaFw0yODA4MTMwNTUyMTBaMDYxNDAyBgNVBAMMK29wZW5z aGlmdC1zZXJ2aWNlLXNlcnZpbmctc2lnbmVyQDE3ODE1MDI3MzAwggEiMA0GCSqG SIb3DQEBAQUAA4IBDwAwggEKAoIBAQDjTXJFMsjKJTQZCjvkQfgeO/XvPMYsbWnF kIFbVtdeBsMrNQ7g80EAkL2J+rA1lWV5zmCKGVJG+tVHviHj8U7wV2YUo+W/Z+/h zCfwE7CBCQH6RUMhv1bb1KDHrPTHiwrc0Y1pg8scjtuxfhH4c/OHBzTYU8wq+/+A FBsrueltXJwq+sTFk6kfqypWYqj0KxM6T3iWACyAxnFX+5aBbXtjhQNY4SphhQql jQA+dH/dhz/z32AvKDxT5jv9z4mpQziB6LggJW9qWqMYVoRFKKb0FaScftZ7j4lh 4IvUyLqTkhWt4ebS4kkXMdoZFJ7fMtaWjuRKhY+0/MBEt5l4JyPtAgMBAAGjYzBh MA4GA1UdDwEB/wQEAwICpDAPBgNVHRMBAf8EBTADAQH/MB0GA1UdDgQWBBTaOuPH Zy/1pBaOl044n4tzc99fwjAfBgNVHSMEGDAWgBTaOuPHZy/1pBaOl044n4tzc99f wjANBgkqhkiG9w0BAQsFAAOCAQEAnyGfVnvkD5OY3rsba7OASXO4qTzitGuImjQz 9zFa2xLLaNqrXWK/e1Kawi6D/DBuZTcWFwAScWHBV+0SQo9Q35x6mYfPcIzPmJig 9BZLkCN2vStONv2bCr713Izi9vBHSYzbyBI1QZnGz93AhG8YNZUBI06VywGR04Vz 7mAWuy9SoQ+tVe8179n+piKkK/LYChmz2uzjamqjJjW9frslPJz0bAxgQH/RG78J tk59/7Wvtqg/FNsKCPBdTjtNRkQ0niBGXxjp4/dsnTCQMFfN9JCXUj9iwekpz96Q OSpWxEA3lE3QkxjLcPlhbWjw6fwj+Q5GEDFSf0x+87ipy/uYuA== -----END CERTIFICATE----- kind: ConfigMap metadata: annotations: service.beta.openshift.io/inject-cabundle: "true" creationTimestamp: "2026-06-15T05:54:52Z" managedFields: - apiVersion: v1 fieldsType: FieldsV1 fieldsV1: f:data: {} f:metadata: f:annotations: .: {} f:service.beta.openshift.io/inject-cabundle: {} manager: kube-controller-manager operation: Update time: "2026-06-15T05:54:52Z" - apiVersion: v1 fieldsType: FieldsV1 fieldsV1: f:data: f:service-ca.crt: {} manager: service-ca-operator operation: Update time: "2026-06-15T05:54:52Z" name: openshift-service-ca.crt namespace: kserve resourceVersion: "12061" uid: fb36734f-6492-4f7d-88c8-7deb78e505f3 kind: ConfigMapList metadata: resourceVersion: "23252"