Overview
The pre-trained APIs are a collection of AI services that include Optical Character Recognition (OCR), Speech-to-Text, and Translation. The services are enabled and disabled using the Kubernetes Resource Model (KRM) API. The KRM API is a management API that is optionally used by the AI Platform Admin.
The purpose of this page is to provide information only.
aiplatform.private.gdc.goog/v1
Package v1 contains API Schema definitions for the aiplatform.private.gdc.goog v1 API group
Alerting
Alerting contains the names of the ConfigMaps that are to be processed and transformed into Observability Platform alerting CRs: MonitoringRule and LoggingRule.
Appears in: - Observability
| Field | Description |
|---|---|
monitoringAlertsConfigMap string |
Configmap with embedded MonitoringRule CR to fetch from Remote K8s Cluster and process by transforming into an Observability Platform MonitoringRule CR and deploy to Admin Cluster. |
loggingAlertsConfigMap string |
ConfigMap with embedded LoggingRule CR to fetch from Remote K8s Cluster and process by transforming into an Observability Platform LoggingRule CR and deploy to Admin Cluster. |
AppConfig
AppConfig is the Schema for the AppConfig API. Should have one per namespace
Appears in: - AppConfigList
| Field | Description |
|---|---|
apiVersion string |
aiplatform.private.gdc.goog/v1 |
kind string |
AppConfig |
metadata ObjectMeta |
Refer to Kubernetes API documentation for fields of metadata. |
spec AppConfigSpec |
|
status AppConfigStatus |
AppConfigList
AppConfigList contains a list of AppConfig
| Field | Description |
|---|---|
apiVersion string |
aiplatform.private.gdc.goog/v1 |
kind string |
AppConfigList |
metadata ListMeta |
Refer to Kubernetes API documentation for fields of metadata. |
items AppConfig array |
AppConfigSpec
Appears in: - AppConfig
| Field | Description |
|---|---|
sql Database |
AppConfigStatus
AppConfigStatus defines the observed state of AppConfig.
Appears in: - AppConfig
| Field | Description |
|---|---|
ready boolean |
Ready indicates whether the resource is in a ready state. |
primaryCondition ResourceCondition |
PrimaryCondition represents the primary condition of a resource. If the resource is ready, then the condition will indicate the resource is ready. Otherwise, the condition will be the primary reason why the resource is not ready. |
resourceConditions ResourceCondition array |
ResourceConditions is a collection of conditions for a resource and its subresources. It can be used to determine the overall health of a resource and its subresources. |
conditions Condition array |
Conditions represents raw resource conditions populated from Kubernetes resources for debugging purpose. |
database DatabaseInfo |
Presents the Database subresource information. |
Dashboards
Dashboards contains a list of the names of Observability Platform-dashboard-compatible configmaps to be processed.
Appears in: - Observability
| Field | Description |
|---|---|
configMaps string array |
ConfigMaps (with embedded Grafana dashboard json definition) to be processed; these are expected to exist in the user cluster and namespace that the AI workloads are deployed to. There should be a 1:1 mapping between the number of desired dashboards and dashboard configmaps. We assume there will be a single key in the configmap: 'data', which will house the dashboard definition json string. |
Database
Database configures the relational db requirements for the target Microservice.
Appears in: - AppConfigSpec
| Field | Description |
|---|---|
postgres PostgresConfig |
DatabaseInfo
DatabaseInfo defines the observed state of Database Information.
Appears in: - AppConfigStatus
| Field | Description |
|---|---|
endpoint string |
|
port string |
HPAStatus
HPAStatus describes the status conditions of the HorizontalPodAutoscaler.
Appears in: - MicroserviceStatus
| Field | Description |
|---|---|
statusConditions HorizontalPodAutoscalerCondition array |
Status conditions |
currentReplicas integer |
Number of current replicas managed by the HPA |
desiredReplicas integer |
Number of desired replicas managed by the HPA |
Logging
Logging configures the log collection functionality for the Observability Platform on GDC air-gapped (do not enable if testing on GKE - this functionality does not apply there).
Appears in: - Observability
| Field | Description |
|---|---|
enableLogCollection boolean |
EnableLogCollection indicates whether or not to enable log collection via the Observability Platform. |
staticLabels object (keys:string, values:string) |
Maps to the 'additionalLabels' property of the OP LoggingTargetCR. Static labels to add to all captured log entries. Label names must not contain dashes ("-"); label values can contain dashes. |
Metrics
Metrics configures how metrics of a pod will be collected.
Appears in: - Observability
| Field | Description |
|---|---|
prometheus PrometheusMetrics |
Microservice
Microservice is the Schema for the services API
Appears in: - MicroserviceList
| Field | Description |
|---|---|
apiVersion string |
aiplatform.private.gdc.goog/v1 |
kind string |
Microservice |
metadata ObjectMeta |
Refer to Kubernetes API documentation for fields of metadata. |
spec MicroserviceSpec |
|
status MicroserviceStatus |
MicroserviceList
MicroserviceList contains a list of Microservice
| Field | Description |
|---|---|
apiVersion string |
aiplatform.private.gdc.goog/v1 |
kind string |
MicroserviceList |
metadata ListMeta |
Refer to Kubernetes API documentation for fields of metadata. |
items Microservice array |
MicroserviceSpec
MicroserviceSpec defines the desired state of Microservice.
Appears in: - Microservice
| Field | Description |
|---|---|
public boolean |
Whether this Microservice is public. If true, it will have an expose DNS record as well as a public endpoint. TLS will be setup on the gateway as well automatically according to the DNS name. The DNS prefix will be the name of the Microservice in metadata. |
enableBucketAccess boolean |
Whether to provision the storage bucket related configurations to the container |
dnsname string |
DNS host name. This will be used to construct DNS host name for the Microservice in GDC environment when Microservice is public |
autoscaling |
Autoscaling configurations. |
observability Observability |
Observability (logging, metrics, monitoring and dashboards) configuration (some of which are applicable only to GDC air-gapped configurations). |
prerequisites Prerequisites |
Prerequisites configuration. |
template PodTemplateSpec |
Template to create pods in the Microservice. The Microservice will contain stateless deployments for these pods. |
PodSpec
PodSpec is a minimalist version of the core PodSpec.
Appears in: - PodTemplateSpec
| Field | Description |
|---|---|
containers Container array |
Specifies containers. Not all attributes are supported. Ports specified from all containers will be exposed. First container (containers[0]) is considered the primary container and will be the one to be tagged with resource requests for node auto-provisioning. The rest of the containers are considered as auxiliary (e.g.,sidecars). |
nodeSelector object (keys:string, values:string) |
NodeSelector is a selector which must be true for the pod to fit on a node. Selector which must match a node's labels for the pod to be scheduled on that node. |
volumes Volume array |
List of volumes that will be mounted to the pods |
initContainers Container array |
List of initialization containers belonging to the pod. Init containers are executed in order prior to containers being started. If any init container fails, the pod is considered to have failed and is handled according to its restartPolicy. The name for an init container or normal container must be unique among all containers. Cannot be updated. More info: https://kubernetes.io/docs/concepts/workloads/pods/init-containers/ |
tolerations Toleration array |
If specified, the pod's tolerations. |
serviceAccount string |
The pre-created Kubernetes service account that will be used by pod. |
PodTemplateSpec
PodTemplateSpec is a container object for PodSpec and potentially metadata spec.
Appears in: - Microservice
| Field | Description |
|---|---|
metadata ObjectMeta |
Refer to Kubernetes API documentation for fields of metadata. |
spec PodSpec |
MicroserviceStatus
MicroserviceStatus defines the observed state of Microservice.
Appears in: - Microservice
| Field | Description |
|---|---|
observedGeneration integer |
The generation observed by the deployment controller. |
endpoint string |
The endpoint that this Microservice is accessible from. Only populated for public services. |
deploymentStatus DeploymentStatus |
The deployment status of this Microservice. |
hpaStatus HPAStatus |
The status of the HorizontalPodAutoscaler. |
ready boolean |
Overall Microservice's readiness status |
conditions Condition array |
Represents the latest available observations of a Microservice's current state. |
Observability
Observability configures the various components and parameters required by the Observability Platform (on GDC air-gapped), or Metrics exposure (on GKE).
Appears in: - MicroserviceSpec
| Field | Description |
|---|---|
metrics Metrics |
Metrics configuration. On GKE these are only collected on the first container. On GDC air-gapped this is used in conjunction with the Observability section to configure the behavior of the Observability Platform. |
logging Logging |
|
alerting Alerting |
|
dashboards Dashboards |
PostgresConfig
PostgresConfig defines the relevant target configuration settings for a Postgres DB.
Appears in: - Database
| Field | Description |
|---|---|
version string |
Prerequisites
Prerequisites configures any prerequisites for the Microservice.
Appears in: - MicroserviceSpec
| Field | Description |
|---|---|
sql boolean |
SQL configuration, if required by the target workload |
PrometheusMetrics
PrometheusMetrics configures the Prometheus metrics exposed by a pod.
Appears in: - Metrics
| Field | Description |
|---|---|
path string |
The metrics URI path to scrape metrics from. Defaults to /metrics. |
port integer |
The port to scrape metrics from. Defaults to 80. |