--- 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@sha256:985e48a4dc8e95274c844c2453803bfeaccfe3a179409acdc978d01063c81ae6", "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@sha256:985e48a4dc8e95274c844c2453803bfeaccfe3a179409acdc978d01063c81ae6", "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.13f9f858-6370-417f-a183-1de0a3348a86.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@sha256:985e48a4dc8e95274c844c2453803bfeaccfe3a179409acdc978d01063c81ae6", "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@sha256:2feaa93986f148e85720c135b637c474e72e04ce9723043ce041d3faf964b128", "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@sha256:91960d110466aedc4a1b6eb167d73f723c2c405bc96bb70d4636000e200b88f8", "memoryLimit": "24Gi", "memoryModelcar": "15Mi", "memoryRequest": "100Mi" } kind: ConfigMap metadata: creationTimestamp: "2026-04-17T11:21:28Z" 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: "2026-04-17T11:21:28Z" - apiVersion: v1 fieldsType: FieldsV1 fieldsV1: f:data: f:ingress: {} manager: kubectl-patch operation: Update time: "2026-04-17T11:22:03Z" name: inferenceservice-config namespace: kserve resourceVersion: "12241" uid: 55283432-9aa7-46f8-aa55-133caf79e84b - apiVersion: v1 data: kserve-agent: quay.io/opendatahub/kserve-agent@sha256:985e48a4dc8e95274c844c2453803bfeaccfe3a179409acdc978d01063c81ae6 kserve-controller: quay.io/opendatahub/kserve-controller@sha256:b35c4a846c2b198dabd1aee0c9222f57fed19f1ff3534deedad8a3fc1c39002f 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-stable 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/kserve-localmodel-controller:latest kserve-localmodelnode-agent: quay.io/opendatahub/kserve-localmodelnode-agent:latest kserve-router: quay.io/opendatahub/kserve-router@sha256:2feaa93986f148e85720c135b637c474e72e04ce9723043ce041d3faf964b128 kserve-storage-initializer: quay.io/opendatahub/kserve-storage-initializer@sha256:91960d110466aedc4a1b6eb167d73f723c2c405bc96bb70d4636000e200b88f8 kube-rbac-proxy: quay.io/opendatahub/odh-kube-auth-proxy@sha256:dcb09fbabd8811f0956ef612a0c9ddd5236804b9bd6548a0647d2b531c9d01b3 llmisvc-controller: ghcr.io/opendatahub-io/kserve/odh-kserve-llmisvc-controller:release-v0.17 kind: ConfigMap metadata: creationTimestamp: "2026-04-17T11:21:28Z" 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: "2026-04-17T11:21:28Z" name: kserve-parameters namespace: kserve resourceVersion: "11849" uid: 6b94142e-06b3-4f79-9ba5-243b059e632c - apiVersion: v1 data: ca.crt: | -----BEGIN CERTIFICATE----- MIIDPDCCAiSgAwIBAgIIDnNGFAHaSMgwDQYJKoZIhvcNAQELBQAwJjESMBAGA1UE CxMJb3BlbnNoaWZ0MRAwDgYDVQQDEwdyb290LWNhMB4XDTI2MDQxNzExMDkzMloX DTM2MDQxNDExMDkzMlowJjESMBAGA1UECxMJb3BlbnNoaWZ0MRAwDgYDVQQDEwdy b290LWNhMIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEA5gnnWM4OxbPf 21FFM9JndLfCOwevbzT9jC1ZouOqv9lStXElQECyzr59BhapNpGkUkLKiBGsPV6s 16TSS+TXE88a+uEfhURVj3rnuHCpzx4AMw+gq6TOXgVUZq+lMbWMXXgJnjGS48RG 8Mk3PRuwDE6bVCA1yySE4/lVgNGITu13/tFFMj/NWORaKF8sZlzCyf7T5HNu36ev 9doBXjBkLpoIVz0ds0mus6UN1YDKXcKXv5hG7atjm97IWVq9Qw8uwRpeq41t0pFp EAXBCDbz5eZp8manUGsLI6S3VV/OscdzcIe9J2Z1qskRDruVXgX5Cj5jISjH/LfN AX2bAvR+XQIDAQABo24wbDAOBgNVHQ8BAf8EBAMCAqQwDwYDVR0TAQH/BAUwAwEB /zBJBgNVHQ4EQgRARi4pKmhlL5CWMyjTvvFuGiec19e4UZw+Z2Wx8OujxbHr+Rod eZD0sRXPXPO9Z5sBTH7pCkLTxO1SzzaqkH7VQzANBgkqhkiG9w0BAQsFAAOCAQEA 5G9ugWVvhw57+SaCFX6VV7CiKbIo8GXXaMkVoSNZFotxZEmKrZUt6LiRyxETRepw LZBWFsmpkwG7rdsI9T5bWLOsplfsd9sZWLYd15ifB1BZhwBM/93K/0m3gZanCM9g VXFn2lHausyUgUqL0TGykZPPqhTOLcc27b5yPumxl1yQkj6aGaq5PBVo6wJZFB7q SbbcgHo8tk0r2dRa3nzwgtrL0cREvFMJ3po+lTZGgzHW2YbL7//vcpzrW6f1Kv5W B04hgGzDnRx5pD0JNay2SlIsc1hc2aDs1pYCpFGHfJHDFvKhnwgOtZfbNHs3V6Ch wbgIdwc99zfS2B/KrrlWhA== -----END CERTIFICATE----- -----BEGIN CERTIFICATE----- MIIEADCCAuigAwIBAgIIU9M+fnqXQXgwDQYJKoZIhvcNAQELBQAwJjESMBAGA1UE CxMJb3BlbnNoaWZ0MRAwDgYDVQQDEwdyb290LWNhMB4XDTI2MDQxNzExMTAxMFoX DTI3MDQxNzExMTAxMFowMDESMBAGA1UEChMJb3BlbnNoaWZ0MRowGAYDVQQDExFv cGVuc2hpZnQtaW5ncmVzczCCASIwDQYJKoZIhvcNAQEBBQADggEPADCCAQoCggEB AL17/L6/PJBQgZcIdKjI8k3F1Z6ImyE7HtIeLu30te9RDSj6RseuBa92zzqpjNK6 wejrrM1SXjKpcjjGLKcVZp7BQBD3E8wfyvrSbupdTl6g+hfjhrWrjRIaDGbSqhw0 RoWDOyEyFbP+zO4dKCpvkngrjYvV2ooif8qUtbrfBYOjHurRlUpEzZ8jX+L9YSny 8Jj/9na7RQKNtfAHYq+j0StiXYXYDuctiuIjkC5cNVoOPl7LC4/NrELlaRKxE8Ht Yz2pvgZ5hCBOvqgbykuj9c80YRfcIHvcHRmrywb2jPAXUfZQFaqiLdQO2DOYWTeD ujk5Fl0imIcUkpILvbLXltcCAwEAAaOCASYwggEiMA4GA1UdDwEB/wQEAwIFoDAd BgNVHSUEFjAUBggrBgEFBQcDAgYIKwYBBQUHAwEwDAYDVR0TAQH/BAIwADBJBgNV HQ4EQgRAgajHivzDNxtCmWWWx5ETEac4xFuRNwx+AcwPrMqJMz+WKFhMN67tteUK oA6A7JQzMT+uhxT9qMIeZt+dw9mVSzBLBgNVHSMERDBCgEBGLikqaGUvkJYzKNO+ 8W4aJ5zX17hRnD5nZbHw66PFsev5Gh15kPSxFc9c871nmwFMfukKQtPE7VLPNqqQ ftVDMEsGA1UdEQREMEKCQCouYXBwcy4xM2Y5Zjg1OC02MzcwLTQxN2YtYTE4My0x ZGUwYTMzNDhhODYucHJvZC5rb25mbHV4ZWFhcy5jb20wDQYJKoZIhvcNAQELBQAD ggEBAFnThJnoWVLQfldiAhschBKxpC/dDkFlfIagAZ3Wztc4tvZ2tAbvHv+CTslM fz9ScGN3oANCZopaF4b6aVgskoBlRnvtnlST374hhlUAO7dcZ94xJi3/dCk+n0zF Vf462O/fv1GdtQh6MBFbaF1xuJNk6RicMBW6L9Gtq0HMNsEJa7/7ndZaVli7gMbH qfHeCoAddAUz4ij/Sw1hMvLobALx0Q8b4T4AqZTwCU/ftS/50iIptAtFJN17wCxE DS4clB1dN4bQsQc3VXmarqbvsJSMOq2r8AGDzBlCfBjOZCVb41/14k26SPqsBXH0 A9QeEDgfnaUspkrFiaAjoRDRQ3w= -----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: "2026-04-17T11:21:14Z" 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: "2026-04-17T11:21:14Z" name: kube-root-ca.crt namespace: kserve resourceVersion: "11665" uid: 37a643b4-5219-4165-99d4-3463ca90bf61 - apiVersion: v1 data: cabundle.crt: |- -----BEGIN CERTIFICATE----- MIIDUTCCAjmgAwIBAgIIbxcnWq2WrzIwDQYJKoZIhvcNAQELBQAwNjE0MDIGA1UE Awwrb3BlbnNoaWZ0LXNlcnZpY2Utc2VydmluZy1zaWduZXJAMTc3NjQyNDU4MjAe Fw0yNjA0MTcxMTE2MjJaFw0yODA2MTUxMTE2MjNaMDYxNDAyBgNVBAMMK29wZW5z aGlmdC1zZXJ2aWNlLXNlcnZpbmctc2lnbmVyQDE3NzY0MjQ1ODIwggEiMA0GCSqG SIb3DQEBAQUAA4IBDwAwggEKAoIBAQDllMoKPY5UYWHs2UhtMpLsRob91rhEosTx WMI9TNQchM0o6ykcsNDyimr0em3PKx9HHPHZRXZKpB79U/Q308WsRjEjRz++Rc05 PcvGa1vQ67ZT9SeMDeq0LT/dUACAg1vHG56V9xXhcKR+onw4H/NrXtqLxWkZfkLK TcHyy7ytt5VqKEDVG7zPmWo1FlbzROyEQ0oEvL2Qk4cBlVDIqfPTI274MuttftYw zEg+m2tjuS79PYg1BVqu5vAFXc7ElPYn0McdHZ/16SHEYA/4p+i+tUUvmZBNmZ2B ctMT50NCrLsz7+hG54ti5T9LZZJyyTSHHfq4O+OmAY3NHeny9r8lAgMBAAGjYzBh MA4GA1UdDwEB/wQEAwICpDAPBgNVHRMBAf8EBTADAQH/MB0GA1UdDgQWBBR3ERzf P/zb9L3VfkbVE+uuj43SxzAfBgNVHSMEGDAWgBR3ERzfP/zb9L3VfkbVE+uuj43S xzANBgkqhkiG9w0BAQsFAAOCAQEAaMVgbPmHuhgx8p1gyjA8iakAq7a2UyJbk+iE 70UxzcIsVn22kbdaSDj5qqcKzmRfzXbFzp04HL8dTSYv06GP79aAPctkFs6GuI3W i801A7V49IK9GYWyAsXYyvkvd9lYEid2oouiwxIajbJwuit333i9Fojy8YIc5xEK rKsVfIcAWj7eecxN3HtYdi7XHkIszg3m1dcvRVOnY+RF+lR8PQstuNANVL77TiBw X5oWG2Aq1NBPm1XpwXirZ1x/eobf9GiTCQH9fYV7bmO3ZkooLOaUr62a7PNJCFmE 9DF/C5ju3R23Rh9EeVQCyT7qnDTWL3mQRKoyToNbvVaAbMVBKA== -----END CERTIFICATE----- kind: ConfigMap metadata: creationTimestamp: "2026-04-17T11:22:42Z" 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: "2026-04-17T11:22:42Z" name: odh-kserve-custom-ca-bundle namespace: kserve resourceVersion: "12674" uid: 798a4378-7166-4644-a573-15ebbcdd5342 - 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: "2026-04-17T11:22:37Z" 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: "2026-04-17T11:22:37Z" name: odh-model-controller-parameters namespace: kserve resourceVersion: "12546" uid: 72b05bb0-9052-4e08-957e-fdffc01525bc - apiVersion: v1 data: service-ca.crt: | -----BEGIN CERTIFICATE----- MIIDUTCCAjmgAwIBAgIIbxcnWq2WrzIwDQYJKoZIhvcNAQELBQAwNjE0MDIGA1UE Awwrb3BlbnNoaWZ0LXNlcnZpY2Utc2VydmluZy1zaWduZXJAMTc3NjQyNDU4MjAe Fw0yNjA0MTcxMTE2MjJaFw0yODA2MTUxMTE2MjNaMDYxNDAyBgNVBAMMK29wZW5z aGlmdC1zZXJ2aWNlLXNlcnZpbmctc2lnbmVyQDE3NzY0MjQ1ODIwggEiMA0GCSqG SIb3DQEBAQUAA4IBDwAwggEKAoIBAQDllMoKPY5UYWHs2UhtMpLsRob91rhEosTx WMI9TNQchM0o6ykcsNDyimr0em3PKx9HHPHZRXZKpB79U/Q308WsRjEjRz++Rc05 PcvGa1vQ67ZT9SeMDeq0LT/dUACAg1vHG56V9xXhcKR+onw4H/NrXtqLxWkZfkLK TcHyy7ytt5VqKEDVG7zPmWo1FlbzROyEQ0oEvL2Qk4cBlVDIqfPTI274MuttftYw zEg+m2tjuS79PYg1BVqu5vAFXc7ElPYn0McdHZ/16SHEYA/4p+i+tUUvmZBNmZ2B ctMT50NCrLsz7+hG54ti5T9LZZJyyTSHHfq4O+OmAY3NHeny9r8lAgMBAAGjYzBh MA4GA1UdDwEB/wQEAwICpDAPBgNVHRMBAf8EBTADAQH/MB0GA1UdDgQWBBR3ERzf P/zb9L3VfkbVE+uuj43SxzAfBgNVHSMEGDAWgBR3ERzfP/zb9L3VfkbVE+uuj43S xzANBgkqhkiG9w0BAQsFAAOCAQEAaMVgbPmHuhgx8p1gyjA8iakAq7a2UyJbk+iE 70UxzcIsVn22kbdaSDj5qqcKzmRfzXbFzp04HL8dTSYv06GP79aAPctkFs6GuI3W i801A7V49IK9GYWyAsXYyvkvd9lYEid2oouiwxIajbJwuit333i9Fojy8YIc5xEK rKsVfIcAWj7eecxN3HtYdi7XHkIszg3m1dcvRVOnY+RF+lR8PQstuNANVL77TiBw X5oWG2Aq1NBPm1XpwXirZ1x/eobf9GiTCQH9fYV7bmO3ZkooLOaUr62a7PNJCFmE 9DF/C5ju3R23Rh9EeVQCyT7qnDTWL3mQRKoyToNbvVaAbMVBKA== -----END CERTIFICATE----- kind: ConfigMap metadata: annotations: service.beta.openshift.io/inject-cabundle: "true" creationTimestamp: "2026-04-17T11:21:14Z" 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: "2026-04-17T11:21:14Z" - apiVersion: v1 fieldsType: FieldsV1 fieldsV1: f:data: f:service-ca.crt: {} manager: service-ca-operator operation: Update time: "2026-04-17T11:21:14Z" name: openshift-service-ca.crt namespace: kserve resourceVersion: "11678" uid: c7b80216-3e7f-4bf0-a055-2e57cf9a5d3c kind: ConfigMapList metadata: resourceVersion: "13730"