Title: Microservice Builder Components — UCS Tools Documentation URL Source: https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html Published Time: Thu, 30 Oct 2025 07:23:04 GMT Markdown Content: Microservice Builder Components[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#microservice-builder-components "Link to this heading") ---------------------------------------------------------------------------------------------------------------------------------------------------------------- Deployment[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#deployment "Link to this heading") ---------------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.app.deployment` **Description:** This component specifies the Kubernetes workload deployment type. You can use it for `Deployment`, `StatefulSet`, `DaemonSet`, and other workload types. You must place this component first in the `spec` section of your manifest. **Parameters** | Field | Type | Description | | --- | --- | --- | | apptype | string (enum) (**Mandatory**) Allowed values: `statefull` (StatefulSet) `stateless` (Deployment) `job` (Job) `cronjob` (CronJob) `daemonset` (DaemonSet) `replicaset` (ReplicaSet) `static-pod` (Pod) | Pod Deployment Type | | extraSpecs | Object | Used for setting extra fields for deployment that are not explicitly supported | **Examples** --- spec: - name: deployment type: ucf.k8s.app.deployment parameters: apptype: stateless extraSpecs: revisionHistoryLimit: 4 To combine multiple workload types in a microservice, create separate `spec` sections in separate YAML documents: --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: backend-container type: ucf.k8s.container ... --- - name: frontend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: frontend-container type: ucf.k8s.container ... Container[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#container "Link to this heading") -------------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.container` **Description:** This component adds a container to the deployment workload. **Parameters** | Field | Type | Description | | --- | --- | --- | | image | object (**Mandatory**) | Container image details | | > .repository | string | Container image repository address | | > .tag | string | Container image tag | | > .config_path | string | Path to container builder configuration file | | > .pullPolicy | string(enum) Allowed values: `Always` `Never` `IfNotPresent` | Container image pull policy | | workload | string | Name reference to a workload component (`ucf.appspec.app.workload`) | | command | list of strings | Array of command line strings to be executed | | args | list of strings | List of arguments to the entry point | | env | list | Array of environment variables to be set | | > .[].name | string | Name of the environment variable | | > .[].value | string | Value for the environment variable | | livenessProbe | object | Liveness probe for the container. Refer to [Probe](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#probe-v1-core) | | readinessProbe | object | Readiness probe for the container. Refer to [Probe](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#probe-v1-core) | | startupProbe | object | Startup probe for the container. Refer to [Probe](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#probe-v1-core) | | ports | list | List of container ports to be exposed | | > .[] | object | Details on container port to expose. Refer to [ContainerPort](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#containerport-v1-core) | | resources | object | Container resource requirements. For more information, refer to [Container Resource Management](https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/) | | > .limits | object | Maximum amount of compute resources allowed, mentioned as key-value pairs. | | > .requests | object | Minimum amount of compute resources required, mentioned as key-value pairs. | | volumeMounts | list | List of volumes to mount in the container. | | > .[] | object | Details on the volume to mount. Refer to [VolumeMount](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#volumemount-v1-core) | | securityContext | object | Security configuration for the container. Refer to [SecurityContext](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#securitycontext-v1-core) | **Examples** # ``name`` of the component is set as the container name - name: server type: ucf.k8s.container parameters: image: repository: nvcr.io/nvidia/pytorch tag: 22.04-py3 command: [sh, -c] args: ["python -m http.server 8080"] ports: - containerPort: 8080 name: http volumeMounts: - name: localvol mountPath: /localvol resources: requests: cpu: 100m memory: 128Mi limits: cpu: 250m memory: 512Mi nvidia.com/gpu: 1 livenessProbe: httpGet: port: http readinessProbe: httpGet: port: http securityContext: runAsNonRoot: true Init Container[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#init-container "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.k8s.initcontainer` **Description:** This component adds an initialization container to the deployment workload. **Parameters** | Field | Type | Description | | --- | --- | --- | | image | string (**Mandatory**) | Container image name | | imagePullPolicy | string(enum) Allowed values: `Always` `Never` `IfNotPresent` | Container image pull policy | | command | list of strings | Array of command line strings to be executed | | args | list of strings | List of arguments to the entry point | | env | list | Array of environment variables to be set | | > .[].name | string | Name of the environment variable | | > .[].value | string | Value for the environment variable | | resources | object | Container resource requirements. For more information, refer to [Container Resource Management](https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/) | | > .limits | object | Maximum amount of compute resources allowed, mentioned as key-value pairs. | | > .requests | object | Minimum amount of compute resources required, mentioned as key-value pairs. | | volumeMounts | list | List of volumes to mount in the container. | | > .[] | object | Details on the volume to mount. Refer to [VolumeMount](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#volumemount-v1-core) | | securityContext | object | Security configuration for the container. Refer to [SecurityContext](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#securitycontext-v1-core) | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless # ``name`` of the component is set as the initcontainer name - name: server type: ucf.k8s.initcontainer parameters: image: nvcr.io/nvidia/pytorch:22.04-py3 command: [sh, -c] args: ["python -m http.server 8080"] volumeMounts: - name: localvol mountPath: /localvol resources: requests: cpu: 100m memory: 128Mi limits: cpu: 250m memory: 512Mi nvidia.com/gpu: 1 securityContext: runAsNonRoot: true Volume[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#volume "Link to this heading") -------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.volume` **Description:** This component adds a volume to the deployment workload. **Parameters** Refer to [Volume](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#volume-v1-core) **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless # ``name`` of the component is set as the volume name # ``name`` will be required to mount the volume in a container - name: configmap-volume type: ucf.k8s.volume parameters: configMap: name: my-app-config - name: data type: ucf.k8s.pvc parameters: annotations: helm.sh/resource-policy: "keep" spec: storageClassName: mdx-local-path # omit this line to use default StorageClass provisioner accessModes: [ReadWriteOnce] resources: requests: storage: 50Mi - name: data-volume type: ucf.k8s.volume parameters: persistentVolumeClaim: claimName: ${CHART_NAME}-data # substitue ${CHART_NAME} with manifest.yaml's chartName value - name: my-container type: ucf.k8s.container parameters: ... volumeMounts: - name: configmap-volume mountPath: /tmp/app-configs - name: data-volume mountPath: /data Volume Claim Template[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#volume-claim-template "Link to this heading") -------------------------------------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.volumeClaimTemplate` **Description:** This component adds a volume claim template for **Stateful** workloads. **Parameters** | Field | Type | Description | | --- | --- | --- | | spec | object (**Mandatory**) | Persistent Volume Claim specification | | > .storageClassName | string | Storage class for the PVC | | > .volumeName | string | Volume name for the PVC | | > .accessModes | list of strings Allowed Values: `ReadWriteOnce` `ReadOnlyMany` `ReadWriteMany` `ReadWriteOncePod` | Access mode of the storage requested. Refer to [Access Modes](https://kubernetes.io/docs/concepts/storage/persistent-volumes#access-modes-1) | | > .resources | object | Storage resource requirements | | > .requests | object | Storage resource requests | | > .storage | string | Minimum storage requirement | | > .limits | object | Storage resource limits | | > .storage | string | Minimum storage requirement | | annotations | object | Annotations to add on the PVC resource as key-value pairs | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: statefull statefulSetServiceName: myservice # ``name`` of the component is set as the volume name # ``name`` will be required to mount the volume in a container - name: data-storage type: ucf.k8s.volumeClaimTemplate parameters: annotations: helm.sh/resource-policy: "keep" spec: accessModes: [ReadWriteOnce] storageClassName: mdx-local-path # omit this line to use default StorageClass provisioner resources: requests: storage: 1Gi - name: my-container type: ucf.k8s.container parameters: ... volumeMounts: - name: data-storage mountPath: /data - name: myservice type: ucf.k8s.service parameters: ports: - port: 1000 name: http-api Default Volume Mount[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#default-volume-mount "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.appspec.defaultVolumeMount` **Description:** This component adds a volume mount to all containers and initialization containers in the deployment workload. **Parameters** Refer to [VolumeMount](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#volumemount-v1-core) **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless # Volume ``data-volume`` will be mounted at ``/data`` in all # containers / init-containers in the deployment # (my-container & my-init-container in this example) - name: data-volume-mount type: ucf.appspec.defaultVolumeMount parameters: name: data-volume mountPath: /data - name: data type: ucf.k8s.pvc parameters: annotations: helm.sh/resource-policy: "keep" spec: storageClassName: mdx-local-path accessModes: [ReadWriteOnce] resources: requests: storage: 50Mi - name: data-volume type: ucf.k8s.volume parameters: persistentVolumeClaim: claimName: ${CHART_NAME}-data # substitue ${CHART_NAME} with manifest.yaml's chartName value - name: my-container type: ucf.k8s.container parameters: ... - name: my-init-container type: ucf.k8s.initcontainer parameters: ... Persistent Volume Claim[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#persistent-volume-claim "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.k8s.pvc` **Description:** This component adds a Persistent Volume Claim to the microservice. **Parameters** | Field | Type | Description | | --- | --- | --- | | spec | object (**Mandatory**) | Persistent Volume Claim specification | | > .storageClassName | string | Storage class for the PVC | | > .volumeName | string | Volume name for the PVC | | > .accessModes | list of strings Allowed Values: `ReadWriteOnce` `ReadOnlyMany` `ReadWriteMany` `ReadWriteOncePod` | Access mode of the storage requested. Refer to [Access Modes](https://kubernetes.io/docs/concepts/storage/persistent-volumes#access-modes-1) | | > .resources | object | Storage resource requirements | | > .requests | object | Storage resource requests | | > .storage | string | Minimum storage requirement | | > .limits | object | Storage resource limits | | > .storage | string | Minimum storage requirement | | annotations | object | Annotations to add on the PVC resource as key-value pairs | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless # ``name`` of the component is set as PVC Name # ``name`` will be required to add the PVC as volume to the pod - name: data type: ucf.k8s.pvc parameters: annotations: helm.sh/resource-policy: "keep" spec: storageClassName: mdx-local-path # omit this line to use default StorageClass provisioner accessModes: [ReadWriteOnce] resources: requests: storage: 50Mi - name: data-volume type: ucf.k8s.volume parameters: persistentVolumeClaim: claimName: ${CHART_NAME}-data # substitue ${CHART_NAME} with manifes.yaml's chartName value - name: my-container type: ucf.k8s.container parameters: ... volumeMounts: - name: data-volume mountPath: /data Service[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#service "Link to this heading") ---------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.service` **Description:** This component adds a service resource to the microservice and associates it with the deployment workload. It creates a Kubernetes service resource. **Parameters** | Field | Type | Description | | --- | --- | --- | | ports | list (**Mandatory**) | List of ports to expose in the service | | > .[].port | integer (**Mandatory**) | Port of the service to expose | | > .[].name | string (**Mandatory**) | Name for service port (this is used as endpoint name) | | > .[].targetPort | string or integer | Number or name for target container port | | > .[].nodePort | integer | Port number of the node to expose this port | | > .[].range | integer | Range of ports. 0- will be added added to port/targetPort/nodePort | | type | string (enum) Allowed Values: `ClusterIP` `NodePort` `LoadBalancer` `ExternalName` | Service type | | externalTrafficPolicy | string (enum) Allowed values: `Cluster` `Local` | Route external traffic to node-local or cluster-wide endpoints | | clusterIP | string | Cluster IP address for the service | | nameOverride | boolean | Enable to set service name as set as - | | fullNameOverride | boolean | Enable to set service name as set as | | annotations | object | Annotations to set on the service as key(string)-value(string) pairs | | labels | object | Labels to set on the service as key(string)-value(string) pairs | | extraSpecs | object | Used for setting extra fields for the service that are not explicitly supported | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: "backend-container" type: ucf.k8s.container parameters: image: repository: nvcr.io/nvidia/pytorch tag: 22.04-py3 command: [sh, -c] args: ["python -m http.server 8080"] ports: - containerPort: 8080 name: http - name: backend-service type: ucf.k8s.service parameters: annotations: some-annotation: annotation-value labels: some-label: label-value ports: - port: http # Must match a container port in the workload name: http-api type: NodePort Image Pull Secret[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#image-pull-secret "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.k8s.imagepullsecret` **Description:** This component adds an image pull secret for pulling container images used in the Pod. Ideally, this should not be part of the microservice since the deployment location or deployer may not be known during microservice development. Instead, you can set it from the application. **Parameters** | Field | Type | Description | | --- | --- | --- | | name | string (**Mandatory**) | Image pull secret name | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: ngc-image-pull-secret type: ucf.k8s.imagepullsecret parameters: name: ngc-docker-reg-secret ConfigMap[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#configmap "Link to this heading") -------------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.configmap` **Description:** This component adds a ConfigMap to the microservice. **Parameters** | Field | Type | Description | | --- | --- | --- | | name | string (**Mandatory**) | Name for the configmap resource | | data | object (**Mandatory**) | Data to add in the configmap as key(string)-value(string) pairs | | annotations | object (**Mandatory**) | Annotations to add on the configmap resource as key(string)-value(string) pairs | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless # ``config.json`` will be mounted at ``/tmp/app-configs/config.json`` in the container - name: my-app-config type: ucf.k8s.configmap parameters: name: my-app-config data: config.json: | { "log-level": "DEBUG" } - name: configmap-volume type: ucf.k8s.volume parameters: configMap: name: my-app-config - name: my-container type: ucf.k8s.container parameters: ... volumeMounts: - name: configmap-volume mountPath: /tmp/app-configs Pod Annotations[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#pod-annotations "Link to this heading") -------------------------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.podAnnotations` **Description:** This component adds annotations to pods belonging to a deployment workload. **Parameters** | Field | Type | Description | | --- | --- | --- | | annotations | object (**Mandatory**) | Annotations to add on the pods as key(string)-value(string) pairs | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: pod-annotations type: ucf.k8s.podAnnotations parameters: kubernetes.io/description: Description of the pod behavior Pod Labels[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#pod-labels "Link to this heading") ---------------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.podLabels` **Description:** This component adds labels to pods belonging to a deployment workload. **Parameters** | Field | Type | Description | | --- | --- | --- | | labels | object (**Mandatory**) | Labels to add on the pods as key(string)-value(string) pairs | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: pod-labels type: ucf.k8s.podLabels parameters: pod-type: "backend-pod" Pod Security Context[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#pod-security-context "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.k8s.podSecurityContext` **Description:** This component sets the security configuration at the pod level. **Parameters** Refer to [Pod Security Context](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#podsecuritycontext-v1-core) **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: podSecurityContext type: ucf.k8s.podSecurityContext parameters: runAsGroup: 1000 runAsUser: 1000 Container Restart Policy[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#container-restart-policy "Link to this heading") -------------------------------------------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.restartPolicy` **Description:** This component sets the container restart policy for all containers in the pod. **Parameters** | Field | Type | Description | | --- | --- | --- | | policy | string(enum) (**Mandatory**) Allowed Values: `Always` `Never` `OnFailure` | Container Restart Policy. Only certain values are allowed based on the workload type. | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: restartPolicy type: ucf.k8s.restartPolicy parameters: policy: Always Pod DNS Policy[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#pod-dns-policy "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.k8s.dnsPolicy` **Description:** This component sets the DNS policy for the pod. Refer to [Pod’s DNS Policy](https://kubernetes.io/docs/concepts/services-networking/dns-pod-service/#pod-s-dns-policy). **Parameters** | Field | Type | Description | | --- | --- | --- | | policy | string(enum) (**Mandatory**) Allowed Values: `Default` `ClusterFirst` `ClusterFirstWithHostNet` `None` | Pod DNS Policy | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: dnsPolicy type: ucf.k8s.dnsPolicy parameters: policy: ClusterFirst PVC Retention Policy[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#pvc-retention-policy "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.k8s.persistentVolumeClaimRetentionPolicy` **Description:** This component sets the retention policy for PVCs created using volumeClaimTemplates of a StatefulSet. It requires Kubernetes 1.27+. **Parameters** | Field | Type | Description | | --- | --- | --- | | whenDeleted | string(enum) Allowed Values: `Retain` (Keep the PVC) `Delete` (Delete the PVC) | PVC retention policy when StatefulSet resource is deleted | | whenScaled | string(enum) Allowed Values: `Retain` (Keep the PVC) `Delete` (Delete the PVC) | PVC retention policy when StatefulSet resource is scaled down | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: statefull - name: pvc-retention-policy type: ucf.k8s.persistentVolumeClaimRetentionPolicy parameters: whenDeleted: Delete whenScaled: Retain Pod Management Policy[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#pod-management-policy "Link to this heading") -------------------------------------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.persistentVolumeClaimRetentionPolicy` **Description:** This component sets the pod replica ordering policy for a **StatefulSet**. Refer to [Pod Management Policies](https://kubernetes.io/docs/concepts/workloads/controllers/statefulset.md/#pod-management-policies). **Parameters** | Field | Type | Description | | --- | --- | --- | | policy | string(enum) Allowed Values: `OrderedReady` (Launch Pod replicas Sequentially) `Parallel` (Launch Pod replicas Parallelly) | Pod replica management policy | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: statefull - name: pvc-management-policy type: ucf.k8s.podManagementPolicy parameters: policy: Parallel Job Options[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#job-options "Link to this heading") ------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.k8s.app.job.jobOptions` **Description:** This component sets various parameters for a **Job** workload. Refer to [Job Spec](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#jobspec-v1-batch). **Parameters** | Field | Type | Description | | --- | --- | --- | | backoffLimit | integer | Number of retries before marking job as failed | | completionMode | string(enum) Allowed Values: `NonIndexed` `Indexed` | Method to track completion of pods | | completions | integer | Number of successfully finished pods this job should be run with | | manualSelector | boolean | Enable user specified selectors | | parallelism | integer | Maximum number of pods that run at any time | | selector | object | Labels to query. Should match job count | | suspend | boolean | Whether the job controller should create pods | | ttlSecondsAfterFinished | integer | Time after job completion after which pods will be deleted | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: job - name: job-options type: ucf.k8s.app.job.jobOptions parameters: completionMode: Indexed parallelism: 3 Job Deadline[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#job-deadline "Link to this heading") -------------------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.app.job.activeDeadlineSeconds` **Description:** This component sets the maximum duration that a pod should execute for a **Job** workload. **Parameters** | Field | Type | Description | | --- | --- | --- | | seconds | integer | Deadline for the job in seconds | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: job - name: job-deadline type: ucf.k8s.app.job.activeDeadlineSeconds parameters: seconds: 100 CronJob Schedule[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#cronjob-schedule "Link to this heading") ---------------------------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.app.cronjob.schedule` **Description:** This component sets the cron schedule for pods of **CronJob** workload. **Parameters** | Field | Type | Description | | --- | --- | --- | | schedule | string | Cronjob schedule string | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: cronjob # Execute daily at 1:00 am - name: cronjob-schedule type: ucf.k8s.app.cronjob.schedule parameters: schedule: "1 0 * * *" Restart Pod On Config Changes[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#restart-pod-on-config-changes "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.appspec.restartPodOnConfigChanges` **Description:** This component restarts pods when any configmaps associated with the microservice are updated. It only supports **Deployment**, **StatefulSet**, and **DaemonSet** workload types. Refer to [Restarting Pods on Config Changes](https://docs.nvidia.com/ucf/2.10.0/text/tutorials/06_Restarting_Pods_Config_Changes.html#tutorial-6). **Parameters** | Field | Type | Description | | --- | --- | --- | | addAll | boolean | Add dependency on all configmaps found in the microservice | | addDefault | boolean | Add dependency on default configmaps added to the microservice viz. - scripts, configs, workload-configs, external-files | | configMaps | list of strings | List of names of configmaps to add dependency on | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: cm-dependencies type: ucf.appspec.restartPodOnConfigChanges parameters: # Add dependency on all configmaps detected in the microservice addAll: true # Add dependency on default configmaps for scripts, configs and workload configs # addDefault: true # Add dependency on individual configmaps using the configmap names # configMaps: # - http-server-configs-cm Ingress[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#ingress "Link to this heading") ---------------------------------------------------------------------------------------------------------------- **Type** `ucf.appspec.ingress` **Description:** This component adds an **Ingress** resource for the microservice. **Parameters** | Field | Type | Description | | --- | --- | --- | | enabled | boolean (**Mandatory**) | Ingress controller state (enabled/disabled) | | className | string | Ingress controller class name | | annotations | object (**Mandatory**) | Annotations to set on the ingress resource as key(string)-value(string) pairs | | rules | list (**Mandatory**) | List of Ingress Rules. Refer to [IngressRule](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#ingressrule-v1-networking-k8s-io) | **Examples** --- spec: - name: backend type: ucf.k8s.app.deployment parameters: apptype: stateless - name: backend-container type: ucf.k8s.container parameters: ... ports: - containerPort: 5000 name: http - name: backend-svc type: ucf.k8s.service parameters: ports: - port: 5000 targetPort: http protocol: TCP name: http type: ClusterIP # Route HTTP calls with /api prefix to the backend service - name: ingress type: ucf.appspec.ingress parameters: enabled: true annotations: nginx.ingress.kubernetes.io/proxy-body-size: "0" rules: - http: paths: - path: /api(/|$)(.*) pathType: Prefix backend: service: name: backend-backend-svc port: number: 5000 Application Workload[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#application-workload "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.appspec.app.workload` **Description:** This component associates an application workload with a container. Fields added to a workload component will be set on the container that is associated with the workload. **Parameters** | Field | Type | Description | | --- | --- | --- | | wl_units | integer | Workload units. This translates to replica count in **Deployments** and **StatefulSets** | | wl_resources: Workload resource requirements | | | | > .requests | object | Workload minimum resource requirements mentioned as key-value pairs | | > .limits | object | Workload maximum limits for resources mentioned as key-value pairs | | wl_data | string | Workload data. This is a free-form text section that will be mounted as /opt/workload-config/ available to containers at run-time | | wl_env | list | Workload additional environment variables | | > .[].name | string | Name of the environment variable | | > .[].value | string | Value for the environment variable | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: backend-workload type: ucf.appspec.app.workload parameters: wl_units: 1 wl_resources: limits: nvidia.com/gpu: 1 wl_data: | inputFile: /root/video.mp4 wl_env: - name: WORKLOAD_NAME env: backend-workload - name: backend-container type: ucf.k8s.container parameters: ... workload: backend-workload UCX Configuration[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#ucx-configuration "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.k8s.ucx-config` **Description:** This component adds configuration for UCX components used in the microservice. UCS Tools will add a container to replace the `$egress` placeholders with the secondary network interface IP address for UCX interface. Refer to [UCX Support in UCS Tools](https://docs.nvidia.com/ucf/2.10.0/text/UCS_ucx.html#ucf-ucx). **Parameters** | Field | Type | Description | | --- | --- | --- | | name | string | Name to create the configmap with | | data | string | Configuration data as a string | | annotations | object | Annotations for the configmap as key(string)-value(string) pairs | | updaterImage | string | Container image to update the config file with update placeholder params | **Examples** Refer to [UCX Support in UCS Tools](https://docs.nvidia.com/ucf/2.10.0/text/UCS_ucx.html#ucf-ucx) Custom Metadata[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#custom-metadata "Link to this heading") -------------------------------------------------------------------------------------------------------------------------------- **Type** `ucf.metadata.custom` **Description:** This component adds user-specified custom metadata for the microservice to its UCS MS Specification. **Parameters** No schema is enforced. You can add any metadata as required in YAML format. **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless # Data in this component will be added to MS spec under ``metadata`` field. - name: custom-data type: ucf.metadata.custom parameters: param1: value1 param2: value2 Pod Scheduling Specification[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#pod-scheduling-specification "Link to this heading") ---------------------------------------------------------------------------------------------------------------------------------------------------------- **Type** `ucf.k8s.scheduling` **Description:** This component specifies the pod scheduling control parameters. **Parameters** | Field | Type | Description | | --- | --- | --- | | nodeSelector | object | k8s node selector definition (Refer to [scheduling](https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#scheduling)). | | tolerations | array | k8s toleration list (Refer to [scheduling](https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#scheduling)). | | affinity | object | k8s affinity definition (Refer to [scheduling](https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#scheduling)). | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: my-service-scheduling type: ucf.k8s.scheduling parameters: nodeSelector: kubernetes.io/os: linux tolerations: - key: key1 operator: Exists effect: NoSchedule affinity: nodeAffinity: requiredDuringSchedulingIgnoredDuringExecution: nodeSelectorTerms: - matchExpressions: - key: some-key-1 operator: In values: [some-value-1, some-value-2] - matchFields: - key: status.phase operator: In values: - Running Service Account Specification[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#service-account-specification "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.k8s.serviceAccount` **Description:** This component specifies a service account for a Pod. **Parameters** | Field | Type | Description | | --- | --- | --- | | name | string | name of the service account | | clusterRole | string | the cluster role which the service account binds | | create | boolean | option to create the service account | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: my-service-sa type: ucf.k8s.serviceAccount parameters: name: my-sa clusterRole: edit create: true Pod Monitor Custom Resource Definition[#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#pod-monitor-custom-resource-definition "Link to this heading") ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ **Type** `ucf.crd.podMonitor` **Description:** This component creates a pod monitor. **Parameters** | Field | Type | Description | | --- | --- | --- | | portName | string | Name of the container port for the metrics endpoint | | path | string | HTTP path where the metrics will be available | **Examples** --- spec: - name: backend-deployment type: ucf.k8s.app.deployment parameters: apptype: stateless - name: "abcd-container" type: ucf.k8s.container parameters: image: repository: ubuntu tag: "latest" ports: - containerPort: 1000 # name: http - containerPort: 10002 # name: ms-metrics - name: ms-metrics type: ucf.crd.podMonitor parameters: portName: ms-metrics path: /metrics Links/Buttons: - [#](https://docs.nvidia.com/ucf/2.10.0/text/UCS_msbuilder_components.html#pod-monitor-custom-resource-definition) - [Probe](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#probe-v1-core) - [ContainerPort](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#containerport-v1-core) - [Container Resource Management](https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/) - [VolumeMount](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#volumemount-v1-core) - [SecurityContext](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#securitycontext-v1-core) - [Volume](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#volume-v1-core) - [Access Modes](https://kubernetes.io/docs/concepts/storage/persistent-volumes#access-modes-1) - [Pod Security Context](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#podsecuritycontext-v1-core) - [Pod’s DNS Policy](https://kubernetes.io/docs/concepts/services-networking/dns-pod-service/#pod-s-dns-policy) - [Pod Management Policies](https://kubernetes.io/docs/concepts/workloads/controllers/statefulset.md/#pod-management-policies) - [Job Spec](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#jobspec-v1-batch) - [Restarting Pods on Config Changes](https://docs.nvidia.com/ucf/2.10.0/text/tutorials/06_Restarting_Pods_Config_Changes.html#tutorial-6) - [IngressRule](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.26/#ingressrule-v1-networking-k8s-io) - [UCX Support in UCS Tools](https://docs.nvidia.com/ucf/2.10.0/text/UCS_ucx.html#ucf-ucx) - [scheduling](https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#scheduling)