ActivateUserResponse
Union, One Of:AddonComponent
unsupported_cluster_types
typing_extensions.Annotated[(typing.Optional[typing.List[ClusterType]], Any)]
🔗 RecommendationRecommendations
AddOnComponentSource
Repository URL of the source
Chart of the source
AddonComponentStatus
Health status of the addon
Sync status of the addon
Indicates whether the addon is installed
Ai21KeyAuth
The API key for AI21 authentication
Ai21Model
Name to identify this AI21 model in the UI
The unique identifier for the AI21 model
🔗 ModelTypeSpecify the type of the AI21 model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
Ai21ProviderAccount
The name of the AI21 provider account
List of integrations that are associated with the AI21 provider account
🔗 CollaboratorList of users who have access to this provider account
AiFeaturesSettings
Enable/Disable AI features. Other fields only apply when this is true
AI provider to use for AI features When TrueFoundry is selected, TrueFoundry will manage and provide the AI models for the features.
Alert
AlertConfig
Type
Name - 3 to 35 lower case characters long alphanumeric word, may contain - in between, cannot start with a number
🔗 NotificationTargetForAlertRuleConfigure one or more notification targets where alerts will be sent. Each target specifies a notification channel (email or slack).
🔗 PrometheusAlertRuleDefine one or more alert rules that specify the conditions to monitor, when to trigger alerts, and how they should be handled.
AlertConfigResource
🔗 AlertConfigResourceTypeThe type of resource this alert config belongs to. Can be either an application or a cluster.
FQN of the resource this alert config belongs to.
AmqpInputConfig
AMQP Queue URL of Subscriber
AMQP Queue Name
Wait timeout for long polling.
AmqpMetricConfig
Upper limit of the number of backlog messages the auto-scaler will try to maintain per replica. If you set this number to 10 and have 30 messages in the stream and one replica, the auto-scaler will scale the number of replicas to 3.
AmqpOutputConfig
AMQP Queue URL of Publisher
AMQP Routing Key to publish to.
AMQP Exchange Name
AnthropicKeyAuth
The API key for Anthropic authentication
AnthropicModel
Name to identify this Anthropic model in the UI
The unique identifier for the Anthropic model
🔗 ModelTypeSpecify the type of the Anthropic model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AnthropicProviderAccount
The name of the Anthropic provider account
List of integrations that are associated with the Anthropic provider account
🔗 CollaboratorList of users who have access to this provider account
Application
🔗 RecommendationRecommendations for this application
alerts_summary
typing_extensions.Annotated[(typing.Optional[typing.Dict[(str, typing.Optional[typing.Any])]], Any)]
Summary of alerts for this application
Debug infos for this application
potential_problems
typing_extensions.Annotated[(typing.Optional[typing.List[ApplicationProblem]], Any)]
🔗 ApplicationProblemPotential problems with the application
ApplicationDebugInfo
ApplicationMetadata
ApplicationProblem
ApplicationsCancelDeploymentResponse
Deployment cancelled
ApplicationSet
Name of the application set.
🔗 ApplicationSetComponentsItemArray of components with their specifications.
Template to be used for the application set.
Values to be used to render components for the application set.
Fully qualified name of the workspace
Flag to indicate if the template manifest should be converted to TrueFoundry manifest
ApplicationSetComponentsItem
Union, One Of:ApplyMlEntityResponse
ApplyMlEntityResponseData
Union, One Of:ApplyTeamRequestManifest
Union, One Of:Artifact
ArtifactManifest
Name of the entity
Key value metadata. Should be valid JSON. For e.g.
\{"business-unit": "sales", "quality": "good", "rating": 4.5\}Version alias is alternate, ideally human readable, version string to reference an artifact version. It should start with
v followed by alphanumeric and it can include . and - in between (e.g. v1.0.0, v1-prod, v3-dev, etc)Name of the ML Repo
Version of the entity
Step/Epoch number in an iterative training loop the artifact version was created. Generally useful when logging a model version from a MLRepo Run
ID of the MLRepo Run that generated the artifact version
ArtifactManifestSource
Union, One Of:ArtifactPath
Local file or folder path
Relative path where the file or folder will be uploaded to in the artifact
ArtifactsCacheVolume
Storage class of the Volume where artifacts will be cached
Size of the Volume (in GB) where artifacts will be cached. Should be greater than twice the size of artifacts getting cached
ArtifactsDownload
🔗 ArtifactsDownloadArtifactsItemList of artifacts to be cached
ArtifactsDownloadArtifactsItem
Union, One Of:ArtifactVersion
AssistantMessage
Role of the message
🔗 AssistantMessageContentText content for the assistant message
🔗 ToolCallList of tool calls made by the assistant
Name of the assistant this message is from
AssistantMessageContent
Union, One Of:AssistantMessageContentItem
Union, One Of:AsyncProcessorSidecar
URL for the processor to invoke
Timeout for the invoke request in seconds
Image for the processor sidecar (This field will be deprecated in the future)
AsyncService
🔗 AsyncServiceReplicasDeploy multiple instances of your pods to distribute incoming traffic across them, ensuring effective load balancing.
AsyncServiceAutoscaling
🔗 AsyncServiceAutoscalingMetricsMetrics to use for the autoscaler
AsyncServiceAutoscalingMetrics
Union, One Of:AsyncServiceReplicas
Union, One Of:AutoRotate
Rotation Interval in days after which the token will be rotated. Minimum value is 30.
Grace Period in days for which the token will be valid after rotation interval. Minimum value is 1.
Autoshutdown
The period to wait after the last received request before scaling the replicas to 0. This value should be high enough to allow for the replicas of the service to come up to avoid premature scaling down.
AwsAccessKeyAuth
AWS Access Key ID
AWS Secret Access Key for the user to authenticate with
AWS Session Token, only required when using temporary credentials
AwsAccessKeyBasedAuth
AWS access key ID that you want to use for authentication.
AWS access key secret that you want to use for authentication.
AwsAssumedRoleBasedAuth
ARN of the role you want to assume.
AwsBedrockGuardrailConfig
The name of the Guardrail Config.
🔗 AwsBedrockGuardrailConfigOperationThe operation type to use for the Guardrail. Validate guardrails are used to validate requests and mutate can validate as well as mutate requests. Validate guardrails are run in parallel while mutate guardrails are run sequentially.
🔗 AwsBedrockGuardrailConfigAuthDataAuthentication data for the AWS account
The ID of the Guardrail to use.
The version of the Guardrail to use.
AwsBedrockGuardrailConfigAuthData
Union, One Of:AwsBedrockProviderAccount
The name of the AWS Bedrock provider account
🔗 AwsBedrockProviderAccountAuthDataAuthentication data for the AWS account
🔗 BedrockModelV2List of integrations that are associated with the AWS Bedrock provider account
🔗 CollaboratorList of users who have access to this provider account
AwsBedrockProviderAccountAuthData
Union, One Of:AwsEcr
The name of the integration that will be displayed in the TrueFoundry UI.
The URL of the registry.
🔗 AwsEcrAuthDataCustom authentication data for the integration.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AwsEcrAuthData
Union, One Of:AwsEksIntegration
The name of the integration that will be displayed in the TrueFoundry UI.
Name of the cluster given in the cloud account.
🔗 AwsEksIntegrationAuthDataCustom authentication data for the integration.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AwsEksIntegrationAuthData
Union, One Of:AwsInferentia
Name of the AWS Inferentia Accccelerator. One of [INF1, INF2]. This field is required for Node Selector and can be ignored in Nodepool Selector.
Count of Inferentia accelerator chips to provide to the application
AwsIntegrations
Union, One Of:AwsParameterStore
The name of the integration that will be displayed in the TrueFoundry UI.
🔗 AwsParameterStoreAuthDataCustom authentication data for the integration.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AwsParameterStoreAuthData
Union, One Of:AwsProviderAccount
The name of the AWS provider account.
The AWS account ID.
🔗 AwsProviderAccountAuthDataAuthentication data for the AWS account.
🔗 AwsIntegrationsList of integrations that are associated with the AWS provider account.
AwsProviderAccountAuthData
Union, One Of:AwsS3
The name of the integration that will be displayed in the TrueFoundry UI.
The root path of the storage.
🔗 AwsS3AuthDataCustom authentication data for the integration.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AwsS3AuthData
Union, One Of:AwsSagemakerProviderAccount
The name of the Sagemaker provider account
🔗 AwsSagemakerProviderAccountAuthDataAuthentication data for the AWS account
🔗 SagemakerModelList of integrations that are associated with the Sagemaker provider account
🔗 CollaboratorList of users who have access to this provider account
AwsSagemakerProviderAccountAuthData
Union, One Of:AwsSecretsManager
The name of the integration that will be displayed in the TrueFoundry UI.
🔗 AwsSecretsManagerAuthDataCustom authentication data for the integration.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AwsSecretsManagerAuthData
Union, One Of:AzureAiInferenceModel
Name to identify this Azure AI model
🔗 ModelTypeTypes of models supported by this Azure AI deployment
deployment_details
typing_extensions.Annotated[(AzureAiInferenceModelDeploymentDetails, Any)]
required
🔗 AzureAiInferenceModelDeploymentDetailsConfiguration details for the Azure AI deployment
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AzureAiInferenceModelDeploymentDetails
Union, One Of:AzureAiManagedDeployment
Name of the Azure AI deployment
Name of the Azure AI endpoint
AzureAiServerlessDeployment
Name of the Azure AI deployment
Region where the Azure AI deployment is located
AzureAksIntegration
The name of the integration that will be displayed in the TrueFoundry UI.
The resource group of the cluster in the cloud account.
Name of the cluster given in the cloud account.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AzureBasicAuth
The username for the Azure managed identity.
The password for the Azure managed identity.
AzureBlobStorage
The name of the integration that will be displayed in the TrueFoundry UI.
The root path of the storage.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AzureConnectionStringAuth
The connection string for the Azure managed identity.
AzureContainerRegistry
The name of the integration that will be displayed in the TrueFoundry UI.
The URL of the registry.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AzureContentSafetyGuardrailConfig
The name of the Guardrail Config.
The resource name where API is deployed.
API version for the Content Safety API
Custom host for the PII detection API
Array of blocklist names to check against
Severity threshold for the Content Safety API
🔗 AzureContentSafetyCategoryCategories to check against
AzureFoundryModel
Name to identify this Azure AI Foundry model
This is the Model Name on Azure
🔗 ModelTypeTypes of models supported by this Azure AI Foundry deployment
Azure Foundry endpoint
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AzureFoundryModelV2
Name to identify this Azure Foundry model in the UI
🔗 ModelTypeTypes of models supported by this Azure AI Foundry deployment
The name of the Azure Foundry deployment
The Azure AI Foundry endpoint URL
The API version for the Azure Foundry model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AzureFoundryProviderAccount
The name of the Azure Foundry provider account
🔗 AzureFoundryModelV2List of integrations that are associated with the Azure Foundry provider account
🔗 CollaboratorList of users who have access to this provider account
AzureIntegrations
Union, One Of:- 🔗 AzureAksIntegration
- 🔗 AzureContainerRegistry
- 🔗 AzureBlobStorage
- 🔗 AzureOpenAiModel
- 🔗 AzureVault
- 🔗 AzureReposIntegration
- 🔗 AzureAiInferenceModel
- 🔗 AzureFoundryModel
AzureKeyAuth
The API key for Azure AI Foundry authentication
AzureOAuth
The Azure Active Directory tenant (directory) ID.
The client (application) ID of an App Registration in the tenant.
A client secret that was generated for the App Registration.
The ID of the target subscription. The value must be an UUID.
AzureOpenAiModel
Display Name - 2 to 62 characters long alphanumeric word, may contain - or . in between, cannot start with a number
This is the name of the deployment
Azure OpenAI Service endpoint
🔗 ModelTypeSpecify the type of the model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AzureOpenAiModelV2
Name to identify this Azure OpenAI model in the UI
The name of the Azure OpenAI model deployment (e.g. gpt-35-turbo, gpt-4). Used for cost tracking and management
The Azure OpenAI API version to use
🔗 ModelTypeSpecify the type of the Azure OpenAI model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AzureOpenAiProviderAccount
The name of the Azure OpenAI provider account
The Azure OpenAI Service endpoint URL
🔗 AzureOpenAiModelV2List of integrations that are associated with the Azure OpenAI provider account
🔗 CollaboratorList of users who have access to this provider account
AzurePiiGuardrailConfig
The name of the Guardrail Config.
The resource name where API is deployed.
API version for the Content Safety API
Custom host for the PII detection API
🔗 AzurePiiGuardrailConfigDomainDomain for PII detection
🔗 AzurePiiCategoryCategories of PII to detect.
Version of the PII detection model to use
Language code for PII detection
AzureProviderAccount
The name of the Azure provider account.
The ID of the target subscription. The value must be an UUID.
🔗 AzureIntegrationsList of integrations that are associated with the Azure provider account.
AzureReposIntegration
Account Name
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
AzureVault
The name of the integration that will be displayed in the TrueFoundry UI.
Name of key vault name.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
BaseArtifactVersion
BaseArtifactVersionManifest
Union, One Of:BaseAutoscaling
Minimum number of replicas to keep available
Maximum number of replicas allowed for the component.
This is the interval to check each trigger on.
BaseOAuth2Login
client ID for OAuth2.
Client secret or the TrueFoundry secret containing the client secret for OAuth2.
URL for the authorization request
The endpoint to exchange auth code for tokens.
List of scopes to request from the OAuth2 provider.
🔗 BaseOAuth2LoginJwtSourceSource of the JWT token to be used for verification.
BaseService
Name of the service. This uniquely identifies this service in the workspace. > Name can only contain alphanumeric characters and ’-’ and can be atmost 25 characters long
🔗 BaseServiceImageSpecify whether you want to deploy a Docker image or build and deploy from source code
Configure environment variables to be injected in the service either as plain text or secrets. Docs
🔗 BaseServiceMountsItemConfigure data to be mounted to service pod(s) as a string, secret or volume. Docs
Labels
Fully qualified name of the workspace
BaseServiceImage
Union, One Of:BaseServiceMountsItem
Union, One Of:BaseWorkbenchInput
Name of the workbench. This uniquely identifies this workbench in the workspace. > Name can only contain alphanumeric characters and ’-’ and can be atmost 25 characters long
Size of the home directory for the workbench (Persistent Storage)
Configure environment variables to be injected in the service either as plain text or secrets. Docs
🔗 BaseWorkbenchInputMountsItemConfigure data to be mounted to workbench pod(s) as a string, secret or volume. Docs
Fully qualified name of the workspace
BaseWorkbenchInputMountsItem
Union, One Of:BasicAuthCreds
Username for service auth - Username for the user to authenticate with
Password for service auth - Password for the user to authenticate with
BedrockKeyAuth
BedrockModel
Display Name - 2 to 62 characters long alphanumeric word, may contain - or . in between, cannot start with a number
🔗 BedrockModelAuthDataAuth Data
🔗 ModelTypeSpecify the type of the model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
BedrockModelAuthData
Union, One Of:BedrockModelV2
Name to identify this Bedrock model in the UI
The unique identifier for the Bedrock model
🔗 ModelTypeSpecify the type of the Bedrock model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
BitbucketIntegration
Account Name
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
BitbucketProviderAccount
BlobStorageReference
Path to the content in blob storage
🔗 MimeTypeMIME type of the content
BlueGreen
Promote the new release to handle the complete traffic. A manual promotion would be needed if this is disabled
Promote the new release to handle the complete traffic after waiting for these many seconds
BudgetConfig
Name of the budget configuration
🔗 BudgetRuleList of budget rules
BudgetRule
Unique identifier for the rule
Numeric limit value for the budget
BudgetWhen
List of subjects that this rule applies to
List of models that this rule applies to
Metadata key-value pairs that this rule applies to
Build
FQN of the container registry. If you can’t find your registry here, add it through the Integrations page
🔗 BuildBuildSourceSource code location.
🔗 BuildBuildSpecInstructions to build a container image out of the build source
BuildBuildSource
Union, One Of:BuildBuildSpec
Union, One Of:BuildInfo
Status of the build: STARTED = 20, SUCCEEDED = 30, FAILED = 40
Canary
🔗 CanaryStepThese steps would be executed in order to enable shifting of traffic slowly from stable to canary version
CanaryStep
Percentage of total traffic to be shifted to the canary release. The rest will continue to go to the existing deployment
Duration for which to pause the release. The release process will wait for these seconds before proceeding to the next step. If this is not set, the step will pause indefinitely on this step
CerebrasKeyAuth
The API key for Cerebras authentication
CerebrasModel
Name to identify this Cerebras model in the UI
ID of the Cerebras model to use
🔗 ModelTypeSpecify the type of the Cerebras model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
CerebrasProviderAccount
The name of the Cerebras provider account
List of integrations that are associated with the Cerebras provider account
🔗 CollaboratorList of users who have access to this provider account
ChangePasswordResponse
Union, One Of:ChatPromptManifest
Name of the entity
Key value metadata. Should be valid JSON. For e.g.
\{"business-unit": "sales", "quality": "good", "rating": 4.5\}Version alias is alternate, ideally human readable, version string to reference an artifact version. It should start with
v followed by alphanumeric and it can include . and - in between (e.g. v1.0.0, v1-prod, v3-dev, etc)Name of the ML Repo
Version of the entity
🔗 ChatPromptManifestMessagesItemList of messages in the chat conversation, must be non-empty
Variables referenced in messages and that can be replaced when running generation
🔗 ToolSchemaList of tools to be used in the chat prompt
🔗 ChatPromptManifestMcpServersItemA list of MCP servers FQNs or URLs and their tools
🔗 ChatPromptManifestResponseFormatResponse format configuration for structured outputs
🔗 ChatPromptManifestRoutingConfigConfiguration for routing requests to different model targets
🔗 McpServerToolDetailsMapping of tool calls to MCP server integration IDs and tool names
ChatPromptManifestMcpServersItem
Union, One Of:ChatPromptManifestMessagesItem
Union, One Of:ChatPromptManifestResponseFormat
Union, One Of:ChatPromptManifestRoutingConfig
Union, One Of:Cluster
ClusterGateway
Used when there are 2 gateways with same host
ClusterManifest
Name - 3 to 35 lower case characters long alphanumeric word, may contain - in between, cannot start with a number
🔗 ClusterManifestClusterTypeThe cluster type of this cluster
Tag the environment like dev, staging or production. You will need to create environments if you don’t have already.
One or more base URLs, which can be either a wildcard domain (resulting in app URLs like
\<app-name\>-\<workspace-name\>.\<base-domain-URL\>) or a non-wildcard domain (resulting in app URLs like \<base-domain-URL\>/\<app-name\>-\<workspace-name\>) > Read more about routing here and about ‘A’ record and domain mapping here🔗 ClusterManifestMonitoringMonitoring
Cluster Default Registry
🔗 ClusterManifestWorkbenchConfigWorkbench Config
Cluster Integration FQN
Workflow Storage Integration
🔗 NodepoolAdd nodepools that are already created in your cluster. When deploying, applications can choose to schedule from these nodepools.
🔗 ClusterManifestNodeLabelKeysThe node label keys that this cluster supports. Note: You will additionally need to add
truefoundry.com/gpu_type label for GPU-supported node pools. For more information, check out this documentation.🔗 CollaboratorCollaborators who can access this cluster
ClusterManifestMonitoring
Cluster Loki URL
Cluster VictoriaLogs URL
Cluster Prometheus URL
Cluster Kubecost URL
ClusterManifestNodeLabelKeys
The nodepool selector label that this cluster supports
ClusterManifestWorkbenchConfig
The default storage class for the home directory of workbench
ClustersDeleteResponse
Success message
Codeserver
CohereKeyAuth
The API key for Cohere authentication
CohereModel
Name to identify this Cohere model in the UI
The name of the Cohere model to use
🔗 ModelTypeSpecify the type of the Cohere model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
CohereProviderAccount
The name of the Cohere provider account
List of integrations that are associated with the Cohere provider account
🔗 CollaboratorList of users who have access to this provider account
Collaborator
Fully Qualified Name of the subject. eg: user:email or team:teamname
Role ID for the resource
CommonToolsSettings
Enable web search tool Allows model to search the web for information.
Enable code executor tool Allows model to execute code and return the results.
Enable sequential thinking tool Allows model to reason step-by-step to solve complex problems.
Enable vision OCR tool Allows model to extract text from images or PDFs.
Enable web scraping tool. Allows model to scrape content from web pages with intelligent format selection for structured or unstructured data.
ContainerTaskConfig
🔗 ContainerTaskConfigImageSpecify whether you want to deploy a Docker image or build and deploy from source code
Configure environment variables to be injected in the task either as plain text or secrets. Docs
🔗 ContainerTaskConfigMountsItemConfigure data to be mounted to Workflow pod(s) as a volume.
Service Account
ContainerTaskConfigImage
Union, One Of:ContainerTaskConfigMountsItem
Union, One Of:CoreNatsOutputConfig
Output NATS URL
Root subject of output NATS
CpuUtilizationMetric
Percentage of cpu request averaged over all replicas which the autoscaler should try to maintain
CreateMultiPartUploadRequest
CreatePersonalAccessTokenResponse
Personal Access Token
CronMetric
Desired number of replicas during the given interval. Default value is max_replicas.
Cron expression indicating the start of the cron schedule.
* * * * * | | | | | | | | | |___ day of week (0-6) (Sunday is 0) | | | |_____ month (1-12) | | |_______ day of month (1-31) | |_________ hour (0-23) |___________ minute (0-59)Cron expression indicating the end of the cron schedule.
* * * * * | | | | | | | | | |___ day of week (0-6) (Sunday is 0) | | | |_____ month (1-12) | | |_______ day of month (1-31) | |_________ hour (0-23) |___________ minute (0-59)Timezone against which the cron schedule will be calculated, e.g. “Asia/Tokyo”. Default is machine’s local time. https://docs.truefoundry.com/docs/list-of-supported-timezones
CustomBasicAuth
The username for the basic authentication.
The password for the basic authentication.
CustomBearerAuth
The bearer token for the custom bearer authentication.
CustomBlobStorage
The name of the integration that will be displayed in the TrueFoundry UI.
The endpoint of the storage provider.
The root path of the storage.
The region where the storage is located.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
CustomGuardrailConfig
The name of the Guardrail Config.
🔗 CustomGuardrailConfigOperationThe operation type to use for the Guardrail. Validate guardrails are used to validate requests and mutate can validate as well as mutate requests. Validate guardrails are run in parallel while mutate guardrails are run sequentially.
🔗 CustomGuardrailConfigTargetSpecify whether the guardrail should be applied to the request or response. Guardrails with target “Request” can be only used in input guardrails and guardrails with target “Response” can only be used in output guardrails.
The URL of the Guardrail to send a post request to.
🔗 CustomGuardrailConfigAuthDataAuthentication data for the Guardrail Server.
Headers for the Guardrail Server. Forwarded to the Guardrail Server as is. For example:
\{"Authorization": "APIKey \<token\>"\}The config for the Guardrail Server. This is a JSON object that will be sent as a config to Guardrail Server along with the request.
CustomGuardrailConfigAuthData
Union, One Of:CustomHelmRepo
The name of the integration that will be displayed in the TrueFoundry UI.
The URL of the Helm Repo.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
CustomIntegrations
Union, One Of:- 🔗 CustomUsernamePasswordArtifactsRegistry
- 🔗 CustomModel
- 🔗 EmailNotificationChannel
- 🔗 CustomHelmRepo
- 🔗 CustomBlobStorage
- 🔗 CustomJwtAuthIntegration
CustomJwtAuthIntegration
The name of the integration that will be displayed in the TrueFoundry UI.
The base URL of the authentication provider.
The JSON Web Key Set URI for JWT verification.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
CustomModel
2 to 62 characters long alphanumeric word, may contain - in between, cannot start with a number
enter valid https/http URL that should not end with trailing slash
🔗 ModelTypeSpecify the type of the model
🔗 CustomModelAuthDataCustom authentication data for the integration.
Custom headers for the integration. Forwarded to the provider as is. For example:
\{"Authorization": "APIKey \<token\>"\}List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
CustomModelAuthData
Union, One Of:CustomProviderAccount
The name of the provider account.
🔗 CustomIntegrationsList of integrations that are associated with the provider account.
CustomTlsSettings
When set to true, it will reject any connection which is not authorized with the list of supplied CAs.
Add custom CA certificates to use for TLS verification.
CustomUsernamePasswordArtifactsRegistry
The name of the integration that will be displayed in the TrueFoundry UI.
The URL of the registry.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
DatabricksApiKeyAuth
API key for Databricks authentication.
DatabricksModel
Name to identify this Databricks model in the UI
The name of the Databricks model to use
🔗 ModelTypeSpecify the type of the Databricks model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
DatabricksProviderAccount
The name of the Databricks provider account
🔗 DatabricksProviderAccountAuthDataDatabricks authentication credentials
The base URL of your Databricks workspace
List of integrations that are associated with the Databricks provider account
🔗 CollaboratorList of users who have access to this provider account
DatabricksProviderAccountAuthData
Union, One Of:DatabricksServicePrincipalAuth
OAuth secret of the Databricks service principal.
Client ID or Application ID of the Databricks service principal.
DataDirectory
DataDirectoryManifest
Name of the data directory
Name of the ML Repo to create the data directory under
Description for the data directory
Key Value metadata. Should be valid JSON. For e.g.
\{"business-unit": "sales", "quality": "good", "rating": 4.5\}DataDirectoryManifestSource
Union, One Of:DeactivateUserResponse
Union, One Of:DeepinfraKeyAuth
The API key for DeepInfra authentication
DeepinfraModel
Name to identify this DeepInfra model in the UI
The name of the DeepInfra model to use
🔗 ModelTypeSpecify the type of the DeepInfra model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
DeepinfraProviderAccount
The name of the DeepInfra provider account
List of integrations that are associated with the DeepInfra provider account
🔗 CollaboratorList of users who have access to this provider account
DeleteApplicationResponse
Application deletion message
DeleteJobRunResponse
Job run deleted
DeletePersonalAccessTokenResponse
Union, One Of:DeleteSecretGroupResponse
Union, One Of:DeleteTeamResponse
Union, One Of:DeleteUserResponse
Union, One Of:DeleteVirtualAccountResponse
Union, One Of:Deployment
deployment_statuses
typing_extensions.Annotated[(typing.Optional[typing.List[DeploymentStatus]], Any)]
applied_recommendations
typing_extensions.Annotated[(typing.Optional[typing.List[Recommendation]], Any)]
🔗 RecommendationApplied recommendations for this deployment
DeploymentBuild
DeploymentManifest
Union, One Of:- 🔗 Service
- 🔗 AsyncService
- 🔗 Job
- 🔗 Notebook
- 🔗 Codeserver
- 🔗 SshServer
- 🔗 RStudio
- 🔗 Helm
- 🔗 Volume
- 🔗 ApplicationSet
- 🔗 Workflow
- 🔗 SparkJob
- 🔗 Intercept
DeploymentStatus
DeveloperMessage
The role of the messages author, in this case developer.
🔗 DeveloperMessageContentThe contents of the developer message.
An optional name for the participant.
DeveloperMessageContent
Union, One Of:DockerFileBuild
The file path of the Dockerfile relative to project root path.
Build context path for the Dockerfile relative to project root path.
🔗 DockerFileBuildCommandOverride the command to run when the container starts When deploying a Job, the command can be templatized by defining
params and referencing them in command E.g. python main.py --learning_rate \{\{learning_rate\}\}Build arguments to pass to docker build
DockerFileBuildCommand
Union, One Of:DockerhubBasicAuth
DockerhubProviderAccount
3 to 32 lower case characters long alphanumeric word, may contain - in between, cannot start with a number
Account Name
DockerhubRegistry
The name of the integration that will be displayed in the TrueFoundry UI.
Registry URL
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
DynamicVolumeConfig
Volume Type for the volume.
Name of the storage class to be used for the volume.
Size of volume in Gi
Specify the notification channel to send alerts to
List of recipients’ email addresses if the notification channel is Email.
EmailNotificationChannel
The name of the integration that will be displayed in the TrueFoundry UI.
The email address from which the email will be sent.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
EmptyResponse
Union, One Of:Endpoint
Host e.g. ai.example.com, app.truefoundry.com
Path e.g. /v1/api/ml/, /v2/docs/
EnkryptAiGuardrailConfig
The name of the Guardrail Config.
🔗 EnkryptAiGuardrailConfigOperationThe operation type to use for the Guardrail. Validate guardrails are used to validate requests and mutate can validate as well as mutate requests. Validate guardrails are run in parallel while mutate guardrails are run sequentially.
The name of the guardrail to use for the Enkrypt AI Guardrail.
EnkryptAiKeyAuth
The API key for Enkrypt AI authentication
Environment
EnvironmentColor
EnvironmentManifest
Environment Name
Environment Type - Indicates if the environment is for production use
Event
Name of the event
Timestamp when the event was first observed
Timestamp when the event was last observed
🔗 EventInvolvedObjectDetails of the involved object
Type of the event
Number of occurrences of the event
Reason for the event
Message describing the event
Namespace of the event
🔗 EventChartChart of the event
EventChart
EventInvolvedObject
ExternalBlobStorageSource
Type of the source
URI referencing a path in the blob storage bucket linked to the MLRepo
FallbackConfig
Name of the fallback configuration
🔗 FallbackRuleList of fallback rules
FallbackModel
Target model or provider FQN to use as fallback
Optional parameters to override in the fallback request
FallbackRule
Unique identifier for the rule
🔗 FallbackModelList of fallback models to try in sequence
FallbackWhen
List of subjects that this rule applies to
List of models that this rule applies to
Metadata key-value pairs that this rule applies to
HTTP status codes that trigger fallback behavior
FastAiFramework
Type of the framework
FiddlerGuardrailConfig
The name of the Guardrail Config.
Threshold for the guardrail.
FiddlerKeyAuth
API key for the Fiddler account
FileInfo
FlyteLaunchPlan
FlyteLaunchPlanId
FlyteLaunchPlanSpec
FlyteTask
FlyteTaskCustom
FlyteTaskCustomTruefoundry
Union, One Of:FlyteTaskId
FlyteTaskTemplate
FlyteWorkflow
FlyteWorkflowId
FlyteWorkflowTemplate
ForwardAction
Service uri to redirect to. This is the kubernetes fully qualified domain name that the target service can be reached on Eg - For a service called
sample-service in workspace ws this would be `sample-service.ws.svc.cluster.localPort to redirect the service traffic to
Function
Name of the tool/function being called
Arguments passed to the tool/function
FunctionSchema
Name of the function
Description of the function
Parameters schema for the function
Indicates if the function should be called strictly
GatewayConfig
Union, One Of:GatewayConfiguration
GcpApiKeyAuth
API Key
GcpGcr
The name of the integration that will be displayed in the TrueFoundry UI.
The URL of the registry.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
GcpGcs
The name of the integration that will be displayed in the TrueFoundry UI.
The root path of the storage.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
GcpGkeIntegration
The name of the integration that will be displayed in the TrueFoundry UI.
Name of the cluster given in the cloud account.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
GcpGsm
The name of the integration that will be displayed in the TrueFoundry UI.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
GcpIntegrations
Union, One Of:GcpKeyFileAuth
Key File Content
GcpProviderAccount
The name of the GCP provider account.
The project id of the GCP account.
🔗 GcpProviderAccountAuthDataAuthentication data for the GCP account.
🔗 GcpIntegrationsList of integrations that are associated with the GCP provider account.
GcpProviderAccountAuthData
Union, One Of:GcpTpu
Name of the TPU Type. One of -
tpu-v4-podslice (TPU v4, ct4p) - tpu-v5-lite-device (TPU v5e, ct5l) - tpu-v5-lite-podslice (TPU v5e, ct5lp) - tpu-v5p-slice (TPU v5p, ct5p)Topology of the TPU slices. Currently only single-host topology is supported. Please refer to TPUs on GKE docs Allowed Values: -
2x2x1 for tpu-v4-podslice - One of 1x1, 2x2, 2x4 for tpu-v5-lite-device and tpu-v5-lite-podslice - 2x2x1 for tpu-v5p-sliceGeminiModelV2
Name to identify this Gemini model in the UI
The unique identifier for the Gemini model
🔗 ModelTypeSpecify the type of the Gemini model (e.g., chat, text, etc.)
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
GetAlertsResponse
🔗 AlertObject containing alert data grouped by alert name
GetApplicationDeploymentResponse
Deployment
GetApplicationResponse
Application
GetArtifactResponse
GetArtifactVersionResponse
GetAuthenticatedVcsurlResponse
Authenticated URL to access the repository
GetAutoProvisioningStateResponse
State of auto-provisioning for the cluster.
GetChartsResponse
Step size
GetClusterResponse
GetDataDirectoryResponse
GetEnvironmentResponse
🔗 EnvironmentEnvironment
GetEventsResponse
GetJobRunResponse
GetLogsResponse
GetMlRepoResponse
GetModelResponse
GetModelVersionResponse
GetOrCreatePersonalAccessTokenResponse
🔗 VirtualAccountPersonal Access Token data
Personal Access Token
Whether the PAT was created (true) or already existed (false)
GetPromptResponse
GetPromptVersionResponse
GetSecretGroupResponse
🔗 SecretGroupSecret Group
GetSecretResponse
Secret
GetSignedUrLsRequest
GetSignedUrLsResponse
GetSuggestedDeploymentEndpointResponse
Host name
Path
🔗 ClusterGatewayGateway configuration
GetTeamResponse
GetTokenForVirtualAccountResponse
Token for the virtual account
GetUserResourcesResponse
🔗 UserResourceResources where user is a collaborator
GetUserResponse
GetUserTeamsResponse
GetVirtualAccountResponse
🔗 VirtualAccountVirtual Account
Virtual Account token (present only when creating a virtual account)
GetWorkspaceResponse
🔗 WorkspaceWorkspace
GitHelmRepo
Git repository URL - Needs to be a valid URL.
Branch/Commit SHA/Tag of the git repo.
Path to the chart.
Helm values files for overriding values in the helm chart. The path is relative to the Path directory defined above
GithubIntegration
Account Name
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
GithubProviderAccount
GitlabIntegration
Account Name
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
GitlabProviderAccount
GitRepositoryExistsResponse
The flag if repository exists or not
The integration id of the repository
GitSource
The repository URL.
The commit SHA.
Selecting branch will select latest commit SHA of the branch.
GluonFramework
Type of the framework
GoogleGeminiProviderAccount
The name of the Google Gemini provider account
🔗 GeminiModelV2List of integrations that are associated with the Google Gemini provider account
🔗 CollaboratorList of users who have access to this provider account
GoogleModel
Display Name - 2 to 62 characters long alphanumeric word, may contain - or . in between, cannot start with a number
🔗 ModelTypeSpecify the type of the model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
GoogleVertexProviderAccount
The name of the Google Vertex provider account
The Google Cloud project ID where Vertex AI is enabled
🔗 VertexModelV2List of integrations that are associated with the Google Vertex provider account
🔗 CollaboratorList of users who have access to this provider account
Graph
Name
Description
Chart name
🔗 GraphChartTypeChart type
Chart params
GroqKeyAuth
The API key for Groq authentication
GroqModel
Name to identify this Groq model in the UI
The name of the Groq model to use
🔗 ModelTypeSpecify the type of the Groq model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
GroqProviderAccount
The name of the Groq provider account
List of integrations that are associated with the Groq provider account
🔗 CollaboratorList of users who have access to this provider account
GuardrailConfigGroup
The name of the Guardrail Config Group.
🔗 CollaboratorList of users who have access to this Guardrail Config Group.
🔗 GuardrailConfigIntegrationsList of Guardrail Configs, which are part of this Guardrail Config Group.
GuardrailConfigIntegrations
Union, One Of:- 🔗 OpenAiModerationsGuardrailConfig
- 🔗 AwsBedrockGuardrailConfig
- 🔗 CustomGuardrailConfig
- 🔗 AzurePiiGuardrailConfig
- 🔗 AzureContentSafetyGuardrailConfig
- 🔗 EnkryptAiGuardrailConfig
- 🔗 PaloAltoPrismaAirsGuardrailConfig
- 🔗 PromptFooGuardrailConfig
- 🔗 FiddlerGuardrailConfig
- 🔗 PangeaGuardrailConfig
- 🔗 PatronusGuardrailConfig
Guardrails
List of guardrail names to apply to user input messages before processing.
List of guardrail names to apply to AI response messages before returning to user.
GuardrailsConfig
Name of the guardrails configuration
🔗 GuardrailsRuleList of guardrail rules
GuardrailsRule
Unique identifier for the rule
Guardrail selectors of the guardrails to apply to the input (format: groupName/guardrailName)
Guardrail selectors of the guardrails to apply to the output (format: groupName/guardrailName)
GuardrailsWhen
List of subjects that this rule applies to
List of models that this rule applies to
Metadata key-value pairs that this rule applies to
H2OFramework
Type of the framework
HeaderMatch
Header name to match on
Header value to match on
HealthProbe
Time to wait after container has started before checking the endpoint
How often to check the endpoint
Time to wait for a response from the endpoint before considering it down
Number of successful responses from the endpoint before container is considered healthy
Number of consecutive failures before the container is considered down
Helm
Name of the Helm deployment. This will be set as the release name of the chart you are deploying.
Add labels to base argo app
🔗 HelmSourceSource helm repository
Values file as block file
ignore_differences
typing_extensions.Annotated[(typing.Optional[typing.List[typing.Dict[(str, typing.Optional[typing.Any])]]], Any)]
Fully qualified name of the workspace
HelmRepo
Helm repository URL - Needs to be a valid URL.
FQN of the helm repo integration. If you can’t find your integration here, add it through the Integrations page
The helm chart name
Helm chart version
HelmSource
Union, One Of:HttpError
HTTP Status Code
Error Message
🔗 HttpErrorCodeError code
Error details
HttpErrorCode
Union, One Of:HttpProbe
Request Type
Path to the health check endpoint
Listening port for the health check endpoint
Host name to connect to, defaults to the pod IP
Scheme to use for connecting to the host
HttpValidationError
HuggingfaceArtifactSource
Model ID of the artifact to be downloaded
Revision of the artifact to be downloaded
List of patterns to ignore while downloading the artifact
Environment variable which will contain the download path of the artifact
IChange
old_value
typing_extensions.Annotated[(typing.Optional[typing.Dict[(str, typing.Optional[typing.Any])]], Any)]
Image
The image URI. Specify the name of the image and the tag. If the image is in Dockerhub, you can skip registry-url (for e.g.
tensorflow/tensorflow). You can use an image from a private registry using Advanced fieldsFQN of the container registry. If you can’t find your registry here, add it through the Integrations page
🔗 ImageCommandOverride the command to run when container starts. When deploying a Job, the command can be templatized by defining
params and referencing them in command E.g. python main.py --learning_rate \{\{learning_rate\}\}ImageCommand
Union, One Of:ImageContentPart
Type of the content part
🔗 ImageContentPartImageUrlImage URL linking to the image
ImageContentPartImageUrl
🔗 ImageContentPartImageUrlUrlImage URL linking to the image
Details
ImageContentPartImageUrlUrl
Union, One Of:InfraProviderAccount
Union, One Of:- 🔗 AwsProviderAccount
- 🔗 AzureProviderAccount
- 🔗 GcpProviderAccount
- 🔗 DockerhubProviderAccount
- 🔗 BitbucketProviderAccount
- 🔗 CustomProviderAccount
- 🔗 GithubProviderAccount
- 🔗 GitlabProviderAccount
- 🔗 JfrogProviderAccount
- 🔗 TtlProviderAccount
- 🔗 TrueFoundryProviderAccount
- 🔗 QuayProviderAccount
- 🔗 SlackProviderAccount
- 🔗 WebhookProviderAccount
- 🔗 PagerDutyProviderAccount
IngressControllerConfig
Ingress Class Name
Whether TLS is managed by the ingress controller. If enabled, the ingress object will have TLS configuration.
InputOutputBasedCostMetricValue
Cost for reading cached tokens (e.g., OpenAI cached prompts, Anthropic cache reads)
Cost for writing/creating cached tokens (e.g., Anthropic cache creation)
Intercept
Name of the service this intercept is for
InterceptRulesItem
Port of service to be intercepted
🔗 InterceptRulesItemActionAction to take on succesful match
InterceptRulesItemAction
Union, One Of:InternalArtifactVersion
InternalListArtifactVersionsResponse
InternalListArtifactVersionsResponseDataItem
Union, One Of:InternalModelVersion
InviteUserResponse
IsClusterConnectedResponse
Whether the cluster is connected
JfrogArtifactsRegistry
The name of the integration that will be displayed in the TrueFoundry UI.
The URL of the registry.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
JfrogBasicAuth
The username for the basic authentication.
The password for the basic authentication.
JfrogProviderAccount
The name of the provider account.
The name of the JFrog account.
List of integrations that are associated with the provider account.
Job
Name of the job
🔗 JobImageSpecify whether you want to deploy a Docker image or build and deploy from source code
🔗 JobTriggerSpecify the trigger
Trigger the job after deploy immediately
🔗 ParamConfigure params and pass it to create different job runs
Configure environment variables to be injected in the service either as plain text or secrets. Docs
🔗 JobAlertConfigure alerts to be sent when the job starts/fails/completes
Specify the maximum number of attempts to retry a job before it is marked as failed.
Job timeout in seconds.
Number of runs that can run concurrently
🔗 JobMountsItemConfigure data to be mounted to job pod(s) as a string, secret or volume. Docs
Labels
Fully qualified name of the workspace
JobAlert
Specify the notification channel to send alerts to
List of recipients’ email addresses if the notification channel is Email.
Send an alert when the job starts
Send an alert when the job fails
JobImage
Union, One Of:JobMountsItem
Union, One Of:JobRun
JobRun Name
Application Name
Deployment Version
Created At
End Time of JobRun
Duration of JobRun
Command
Total Retries
Error
🔗 JobRunStatusStatus of JobRun
Triggered By
🔗 SubjectTriggered By Subject
Exit Code
Spark UI Url
JobTrigger
Union, One Of:JobTriggerInput
🔗 JobTriggerInputCommandCommand
Parameters used to populate template command
JobTriggerInputCommand
Union, One Of:JsonObjectResponseFormat
Type of response format
JsonSchema
Name of the schema
JSON schema object defining the expected structure
JsonSchemaResponseFormat
Type of response format
Jwt
JwtAuthConfig
FQN of the JWT Auth integration. If you can’t find your integration here, add it through the Integrations page
Enable login for the service
🔗 JwtAuthConfigClaimsItemList of key-value pairs of claims to verify in the JWT token
List of paths that will bypass auth. needs to start with a forward slash(/) and should not contain wildcards(*)
JwtAuthConfigClaimsItem
KafkaInputConfig
‘Kafka Bootstrap servers - Comma separated list of Kafka brokers “hostname:port” to connect to for bootstrap’
Kafka topic to subscribe to
The name of the consumer group to join for dynamic partition assignment
TLS configuration for SASL authentication
Wait timeout for long polling.
KafkaMetricConfig
Upper limit of the number of backlog messages the auto-scaler will try to maintain per replica. If you set this number to 10 and have 30 messages in the stream and one replica, the auto-scaler will scale the number of replicas to 3.
KafkaOutputConfig
‘Kafka Bootstrap servers - Comma separated list of Kafka brokers “hostname:port” to connect to for bootstrap’
Kafka topic to publish to
TLS configuration for SASL authentication
KafkaSaslAuth
Username for SASL authentication
Password for SASL authentication
KerasFramework
Type of the framework
Kustomize
Content of kustomization.yaml to perform kustomize operation. Please do not include the
resources section. It is filled in automaticallyAdditional kubernetes manifests to be included in the application
LatencyBasedLoadBalanceTarget
Target model or provider FQN
Status Codes for which the request will fallback to other targets. If the status code is not present in fallback_status_codes, it fails immediately.
Whether this target is a fallback candidate. If set to false, this model will not be considered as a fallback option for targets of this load-balance-rule
Optional parameters to override in the request
LatencyBasedLoadBalancing
Routing Type
🔗 LatencyBasedLoadBalanceTargetList of targets for latency-based load balancing
LatencyBasedLoadBalancingRule
Unique identifier for the rule
LightGbmFramework
Type of the framework
ListApplicationDeploymentsResponse
Array of Deployments
🔗 PaginationPagination information
ListApplicationsResponse
Array of Applications
🔗 PaginationPagination information
ListArtifactsResponse
ListArtifactVersionsResponse
ListClusterAddonsResponse
🔗 AddonComponentArray of Cluster Addons
🔗 PaginationPagination information
ListClustersResponse
🔗 PaginationPagination information
ListDataDirectoriesResponse
ListEnvironmentsResponse
🔗 EnvironmentArray of Environments
🔗 PaginationPagination information
ListFilesRequest
ListFilesResponse
ListJobRunResponse
🔗 PaginationPagination Information
ListMlReposResponse
ListModelsResponse
ListModelVersionsResponse
ListPersonalAccessTokenResponse
🔗 VirtualAccountArray of Personal Access Tokens
🔗 PaginationPagination Information
ListPromptsResponse
ListPromptVersionsResponse
ListSecretGroupResponse
🔗 SecretGroupArray of Secret Groups
🔗 PaginationPagination information
ListSecretsResponse
Array of Secrets
🔗 PaginationPagination Information
ListTeamsResponse
🔗 PaginationPagination Information
ListUsersResponse
🔗 PaginationPagination information
ListVirtualAccountResponse
🔗 VirtualAccountArray of Virtual Accounts
🔗 PaginationPagination Information
ListWorkspacesResponse
🔗 WorkspaceArray of Workspace
🔗 PaginationPagination information
LoadBalanceTarget
Target model or provider FQN
Relative weight for routing decisions (higher values mean more traffic)
Status Codes for which the request will fallback to other targets. If the status code is not present in fallback_status_codes, it fails immediately.
Whether this target is a fallback candidate. If set to false, this model will not be considered as a fallback option for targets of this load-balance-rule
Optional parameters to override in the request
LoadBalancingConfig
Name of the load balancing configuration
🔗 LoadBalancingRuleList of load balancing rules
LoadBalancingRule
Union, One Of:LoadBalancingWhen
List of subjects that this rule applies to
List of models that this rule applies to
Metadata key-value pairs that this rule applies to
LocalArtifactSource
Type of the source
🔗 ArtifactPathArray of ArtifactPath objects representing the source and destination paths
LocalModelSource
Type of the source
Path to the model file or folder
LocalSource
Local project root path.
run docker build locally
Log
Job Name
Log Content
Log Stream
Log Time
Container Name
Manual
McpServerAuth
Union, One Of:McpServerHeaderAuth
McpServerHeaderOverrideAuth
McpServerIntegration
The name of the MCP Server.
Provide a brief description of the purpose of this MCP Server.
The endpoint URL for the MCP Server. The system will first try a connection using streamable-http transport on this URL. If that fails, it will attempt a connection using SSE transport on <url>/sse.
List of teams/users who can use this MCP server.
McpServerIntegrations
Union, One Of:McpServerOAuth2
OAuth2 authentication
client ID for OAuth2 or the TrueFoundry secret FQN containing the client ID.
Client secret or the TrueFoundry secret FQN containing the client secret for OAuth2.
URL for the authorization request
The endpoint to exchange auth code for tokens.
List of scopes to request from the OAuth2 provider.
🔗 McpServerOAuth2JwtSourceSource of the JWT token to be used for verification.
McpServerOAuth2Dcr
Uses Dynamic Client Registration (RFC 7591) to automatically obtain OAuth2 credentials from the MCP server. The system will register as a client, receive client ID and secret, and handle the complete OAuth2 flow without manual configuration.
McpServerPassthrough
McpServerProviderAccount
The name of the MCP Server Group.
🔗 CollaboratorList of teams/users who can manage this MCP Server Group. These subjects can add new MCP server integrations, can edit existing ones and can also use the MCP server.
🔗 McpServerIntegrationsList of MCP Servers, which are part of this MCP Server Group.
McpServerToolDetails
The integration ID of the MCP server
The name of the tool in the MCP server
McpServerWithFqn
The FQN of the MCP server
The FQN of the MCP server
Whether to enable all tools from the MCP server
McpServerWithUrl
The URL of the MCP server
The URL of the MCP server
The headers to send to the MCP server
Whether to enable all tools from the MCP server
McpTool
The name of the tool as it appears in the MCP server
Metadata
Job Run Name Alias
Metric
MirrorAction
Service uri to mirror to. This is the kubernetes fully qualified domain name that the target service can be reached on Eg - For a service called
sample-service in workspace ws this would be `sample-service.ws.svc.cluster.localPort to redirect the service traffic to
Percentage of requests to mirror
MistralAiKeyAuth
The API key for Mistral AI authentication
MistralAiModel
Name to identify this Mistral AI model in the UI
The name of the Mistral AI model to use
🔗 ModelTypeSpecify the type of the Mistral AI model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
MistralAiProviderAccount
The name of the Mistral AI provider account
List of integrations that are associated with the Mistral AI provider account
🔗 CollaboratorList of users who have access to this provider account
MlRepo
Tenant Name
Created At
Experiment Id
Number of runs
Number of datasets
MlRepoManifest
Name of the ML Repo.
Description for the ML Repo.
Storage Integration to store artifacts and models. A storage integration represents a connected blob storage like AWS S3 / Azure Blob Storage / Google Cloud Storage.
🔗 CollaboratorUsers and Teams that have access to MLRepo
Model
ModelConfiguration
Name of the provider, must be one of the integration providers configured for the Gateway
Name of the model to be used for generations. This model should be available in the provider
🔗 ParametersParameters to pass to the model when generating
Arbitrary extra parameters to pass to the model when generating
ModelCostMetric
Union, One Of:ModelManifest
Name of the entity
Key value metadata. Should be valid JSON. For e.g.
\{"business-unit": "sales", "quality": "good", "rating": 4.5\}Version alias is alternate, ideally human readable, version string to reference an artifact version. It should start with
v followed by alphanumeric and it can include . and - in between (e.g. v1.0.0, v1-prod, v3-dev, etc)Name of the ML Repo
Version of the entity
🔗 ModelManifestFrameworkFramework for the model version like Transformers, PyTorch, Sklearn, Xgboost etc with framework specific metadata. This will be used to infer model deployment configuration
Step/Epoch number in an iterative training loop the model version was created. Generally useful when logging a model version from a MLRepo Run
ID of the MLRepo Run that generated the model version
ModelManifestFramework
Union, One Of:- 🔗 TransformersFramework
- 🔗 TensorFlowFramework
- 🔗 SklearnFramework
- 🔗 PyTorchFramework
- 🔗 KerasFramework
- 🔗 XgBoostFramework
- 🔗 LightGbmFramework
- 🔗 FastAiFramework
- 🔗 H2OFramework
- 🔗 OnnxFramework
- 🔗 SpaCyFramework
- 🔗 StatsModelsFramework
- 🔗 GluonFramework
- 🔗 PaddleFramework
ModelManifestSource
Union, One Of:ModelProviderAccount
Union, One Of:- 🔗 AwsBedrockProviderAccount
- 🔗 GoogleVertexProviderAccount
- 🔗 GoogleGeminiProviderAccount
- 🔗 AzureOpenAiProviderAccount
- 🔗 AzureFoundryProviderAccount
- 🔗 CohereProviderAccount
- 🔗 Ai21ProviderAccount
- 🔗 AnthropicProviderAccount
- 🔗 DeepinfraProviderAccount
- 🔗 GroqProviderAccount
- 🔗 MistralAiProviderAccount
- 🔗 PerplexityAiProviderAccount
- 🔗 TogetherAiProviderAccount
- 🔗 NomicProviderAccount
- 🔗 PalmProviderAccount
- 🔗 OllamaProviderAccount
- 🔗 OpenaiProviderAccount
- 🔗 DatabricksProviderAccount
- 🔗 SelfHostedModelProviderAccount
- 🔗 OpenRouterProviderAccount
- 🔗 SambaNovaProviderAccount
- 🔗 AwsSagemakerProviderAccount
- 🔗 CerebrasProviderAccount
ModelVersion
ModelVersionEnvironment
Python version to use when running the model
pip dependencies needed for running the model
MultiPartUpload
MultiPartUploadResponse
NatsInputConfig
Input NATS URL
Name of the NATS stream
Root subject of input NATS
Consumer name of input NATS
Wait timeout for long polling.
URL for the NATS metrics endpoint. It is compulsory if you want to use NATS autoscaling.
NatsMetricConfig
Upper limit of the number of backlog messages the auto-scaler will try to maintain per replica. If you set this number to 10 and have 30 messages in the stream and one replica, the auto-scaler will scale the number of replicas to 3.
NatsOutputConfig
Output NATS URL
Root subject of output NATS
NatsUserPasswordAuth
Name of the NATS account
User for NATS authentication
Password for NATS authentication
Nodepool
NodepoolSelector
Nodepools where you want to run your workload. Multiple nodepools can be selected. The workload is guaranteed to be scheduled on one of the nodepool
NodeSelector
Instance family of the underlying machine to use. Multiple instance families can be supplied. The workload is guaranteed to be scheduled on one of them.
🔗 NodeSelectorCapacityTypeConfigure what type of nodes to run the app. By default no placement logic is applied. “spot_fallback_on_demand” will try to place the application on spot nodes but will fallback to on-demand when spot nodes are not available. “spot” will strictly place the application on spot nodes. “on_demand” will strictly place the application on on-demand nodes.
NomicKeyAuth
The API key for Nomic authentication
NomicModel
Name to identify this Nomic model in the UI
The name of the Nomic model to use
🔗 ModelTypeSpecify the type of the Nomic model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
NomicProviderAccount
The name of the Nomic provider account
List of integrations that are associated with the Nomic provider account
🔗 CollaboratorList of users who have access to this provider account
Notebook
Stop the notebook instance after this much time in minutes of inactivity. The notebook instance will be stopped even if the notebook is open in your browser, but nothing is running on the notebook.
NotebookConfig
The base domain for the cluster with which you can access your Notebooks
NotificationTarget
Union, One Of:NotificationTargetForAlertRule
Union, One Of:NvidiaGpu
Name of the Nvidia GPU. One of [P4, P100, V100, T4, A10G, A100_40GB, A100_80GB] This field is required for Node Selector and can be ignored in Nodepool Selector. One instance of the card contains the following amount of memory - P4: 8 GB, P100: 16 GB, V100: 16 GB, T4: 16 GB, A10G: 24 GB, A100_40GB: 40GB, A100_80GB: 80 GB
Count of GPUs to provide to the application Note the exact count and max count available for a given GPU type depends on cloud provider and cluster type.
NvidiaMiggpu
Name of the Nvidia GPU. One of [P4, P100, V100, T4, A10G, A100_40GB, A100_80GB] This field is required for Node Selector and can be ignored in Nodepool Selector. One instance of the card contains the following amount of memory - P4: 8 GB, P100: 16 GB, V100: 16 GB, T4: 16 GB, A10G: 24 GB, A100_40GB: 40GB, A100_80GB: 80 GB
🔗 NvidiaMiggpuProfileName of the MIG profile to use. One of the following based on gpu type Please refer to https://docs.nvidia.com/datacenter/tesla/mig-user-guide/#supported-mig-profiles for more details A30 - [1g.6gb, 2g.12gb, 4g.24gb] A100 40 GB - [1g.5gb, 1g.10gb, 2g.10gb, 3g.20gb, 4g.20gb, 7g.40gb] A100 80 GB / H100 80 GB - [1g.10gb, 1g.20gb, 2g.20gb, 3g.40gb, 4g.40gb, 7g.80gb] H100 94 GB - [1g.12gb, 1g.24gb, 2g.24gb, 3g.47gb, 4g.47gb, 7g.94gb] H100 96 GB - [1g.12gb, 1g.24gb, 2g.24gb, 3g.48gb, 4g.48gb, 7g.96gb] H200 141 GB - [1g.18gb, 1g.35gb, 2g.35gb, 3g.71gb, 4g.71gb] B200 180 GB - [1g.23gb, 1g.45gb, 2g.45gb, 3g.90gb, 4g.90gb, 7g.180gb]
NvidiaTimeslicingGpu
Name of the Nvidia GPU. One of [P4, P100, V100, T4, A10G, A100_40GB, A100_80GB] This field is required for Node Selector and can be ignored in Nodepool Selector. One instance of the card contains the following amount of memory - P4: 8 GB, P100: 16 GB, V100: 16 GB, T4: 16 GB, A10G: 24 GB, A100_40GB: 40GB, A100_80GB: 80 GB
Amount of GPU memory (in MB) to allocate. Please note, this limit is not being enforced today but will be in future. Applications are expected to operate in co-opertative mode
OAuth2LoginProvider
OciRepo
OCI chart URL - Need to be a valid URL.
FQN of the container registry. If you can’t find your registry here, add it through the Integrations page
Helm chart version
OllamaKeyAuth
The custom host URL for Ollama API
OllamaModel
Name to identify this Ollama model in the UI
The name of the Ollama model to use
🔗 ModelTypeSpecify the type of the Ollama model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
OllamaProviderAccount
The name of the Ollama provider account
List of integrations that are associated with the Ollama provider account
🔗 CollaboratorList of users who have access to this provider account
OnnxFramework
Type of the framework
OpenaiApiKeyAuth
Your OpenAI API key for authentication
OpenAiModel
Name to identify this OpenAI model in the UI
The name of the OpenAI model to use
🔗 ModelTypeSpecify the type of the OpenAI model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
OpenAiModerationsGuardrailConfig
The name of the Guardrail Config.
Optional custom base URL for OpenAI API. If not provided, the default base URL will be used.
The model to use for the OpenAI Moderation API.
category_thresholds
typing.Optional[typing.Dict[(str, OpenAiModerationsGuardrailConfigCategoryThresholdsValue)]]
🔗 OpenAiModerationsGuardrailConfigCategoryThresholdsValueThe thresholds for the OpenAI Moderation API.
OpenAiModerationsGuardrailConfigCategoryThresholdsValue
Union, One Of:OpenAiModerationsGuardrailConfigCategoryThresholdsValueHarassment
OpenaiProviderAccount
The name of the OpenAI provider account
Optional custom base URL for OpenAI API
List of integrations that are associated with the OpenAI provider account
🔗 CollaboratorList of users who have access to this provider account
OpenRouterApiKeyAuth
Your OpenRouter API key for authentication
OpenRouterModel
Name to identify this OpenRouter model in the UI
The name of the OpenRouter model to use
🔗 ModelTypeSpecify the type of the OpenRouter model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
OpenRouterProviderAccount
The name of the OpenRouter provider account
List of integrations that are associated with the OpenRouter provider account
🔗 CollaboratorList of users who have access to this provider account
OwnedBy
The name of the team who owns the virtual account
PaddleFramework
Type of the framework
PagerDuty
Specify the notification channel to send alerts to
PagerDutyIntegration
The name of the integration that will be displayed in the TrueFoundry UI.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
PagerDutyIntegrationKeyAuth
The integration key for the PagerDuty integration
The integration URL for the PagerDuty integration
PagerDutyProviderAccount
The name of the PagerDuty provider account
🔗 PagerDutyIntegrationList of integrations that are associated with the PagerDuty provider account
Pagination
Total number of items
Number of items skipped
Number of items per page
PalmKeyAuth
Your PaLM API key for authentication
PalmModel
Name to identify this PaLM model in the UI
The name of the PaLM model to use
🔗 ModelTypeSpecify the type of the PaLM model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
PalmProviderAccount
The name of the PaLM provider account
List of integrations that are associated with the PaLM provider account
🔗 CollaboratorList of users who have access to this provider account
PaloAltoPrismaAirsGuardrailConfig
The name of the Guardrail Config.
The profile name associated with this Guardrail Config.
🔗 PaloAltoPrismaAirsGuardrailConfigModeExecution mode for the guardrail. Sync waits for the guardrail check to complete before proceeding. Async triggers the check without waiting. Defaults to sync.
PaloAltoPrismaAirsKeyAuth
API key for the Palo Alto Prisma AIRS account
PangeaGuardrailConfig
The name of the Guardrail Config.
Domain of the cloud provider and region where your Pangea project is configured. Example: if endpoint is: https://<service_name>.aws.us-west-2.pangea.cloud/v1/text/guard, the input should be: aws.us-west-2.pangea.cloud
Recipe key of a configuration of data types and settings defined in the Pangea User Console. It specifies the rules that are to be applied to the text, such as defang malicious URLs.
Overrides for the Pangea account
PangeaKeyAuth
API key for the Pangea account
Param
Name of the param
Description of param
Default value or placeholder
Parameters
ParametersStop
Union, One Of:PatronusAnswerRelevanceEvaluator
The type of the evaluator to use for the Patronus Guardrail.
PatronusEvaluator
Union, One Of:- 🔗 PatronusAnswerRelevanceEvaluator
- 🔗 PatronusGliderEvaluator
- 🔗 PatronusJudgeEvaluator
- 🔗 PatronusPiiEvaluator
- 🔗 PatronusPhiEvaluator
- 🔗 PatronusToxicityEvaluator
PatronusGliderEvaluator
The type of the evaluator to use for the Patronus Guardrail.
PatronusGuardrailConfig
The name of the Guardrail Config.
🔗 PatronusGuardrailConfigTargetThe type of request to use for the Patronus Guardrail.
🔗 PatronusEvaluatorThe evaluators to use for the Patronus Guardrail.
PatronusJudgeEvaluator
The type of the evaluator to use for the Patronus Guardrail.
PatronusKeyAuth
API key for the Patronus account
PatronusPhiEvaluator
The type of the evaluator to use for the Patronus Guardrail.
PatronusPiiEvaluator
The type of the evaluator to use for the Patronus Guardrail.
PatronusToxicityEvaluator
The type of the evaluator to use for the Patronus Guardrail.
Permissions
The fully qualified name of the resource
The type of the resource (cluster, workspace, etc.)
The role id of the role to be assigned to the service account for that resource
PerplexityAiKeyAuth
Your Perplexity AI API key for authentication
PerplexityAiModel
Name to identify this Perplexity AI model in the UI
The name of the Perplexity AI model to use
🔗 ModelTypeSpecify the type of the Perplexity AI model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
PerplexityAiProviderAccount
The name of the Perplexity AI provider account
List of integrations that are associated with the Perplexity AI provider account
🔗 CollaboratorList of users who have access to this provider account
PersonalAccessTokenManifest
Personal Access Token Name
Expiration Date of the Personal Access Token (should be in the ISO 8601 format)
The fully qualified name of the user
PerThousandEmbeddingTokensCostMetric
PerThousandTokensCostMetric
Pip
Path to
requirements.txt relative to Path to build contextDefine pip package requirements. In Python/YAML E.g. [“fastapi>=0.90,<1.0”, “uvicorn”]
Poetry
Poetry version to use
install options to pass to poetry command
PolicyFilters
Cluster names where Policy applies
Workspace names where Policy applies
Environment names where Policy applies
PolicyManifest
Unique identifier of the policy across the organisation
Description of the policy that explains what the policy does, its purpose, and how it affects resources. This helps other users understand the policy’s behavior and impact.
🔗 PolicyManifestOperationDefines if the policy validates or mutates resources. Validation policies check TrueFoundry manifests to ensure they meet conditions and can block deployments. Mutation policies modify Kubernetes manifests before they’re applied to the cluster. See this documentation for more details.
🔗 PolicyManifestModeMode of the policy:
Audit logs all policy evaluations without blocking deployments. Enforce blocks deployments if the policy fails. Disabled deactivates the policy.🔗 PolicyEntityTypesTypes of applications this policy applies to
Actions this policy applies to
TypeScript code for Policy logic. To write and test your policy code, please refer to this repository
PolicyManifestOperation
Union, One Of:PolicyMutationOperation
Mutates or Validates Resources
Determines the sequence in which mutation policies are executed. Policies with lower order value run first, followed by higher value. For example, a policy with order 10 runs before one with order 20. This is crucial when multiple mutation policies need to run in a specific sequence. The order must be between 1 and 100.
PolicyValidationOperation
Mutates or Validates Resources
Port
Port number to expose.
🔗 PortProtocolProtocol for the port.
Expose the port
🔗 PortAppProtocolApplication Protocol for the port. Select the application protocol used by your service. For most use cases, this should be
http(HTTP/1.1). If you are running a gRPC server, select the grpc option. This is only applicable if expose=true.Host e.g. ai.example.com, app.truefoundry.com
Path e.g. /v1/api/ml/, /v2/docs/
Rewrite the path prefix to a different path. If
path is /v1/api and rewrite_path_to is /api. The URI in the HTTP request http://0.0.0.0:8080/v1/api/houses will be rewritten to http://0.0.0.0:8080/api/houses before the request is forwarded your service. Defaults to /. This is only applicable if path is given.🔗 PortAuthAuthentication method for inbound traffic
PortAuth
Union, One Of:PresignedUrlObject
PriorityBasedLoadBalanceTarget
Target model or provider FQN
Priority for the target, Lower the number, higher the priority (0 is the highest priority)
Status Codes for which the request will fallback to other targets. If the status code is not present in fallback_status_codes, it fails immediately.
Whether this target is a fallback candidate. If set to false, this model will not be considered as a fallback option for targets of this load-balance-rule
Optional parameters to override in the request
PriorityBasedLoadBalancing
Routing Type
🔗 PriorityBasedLoadBalanceTargetList of targets for priority-based load balancing
PriorityBasedLoadBalancingRule
Unique identifier for the rule
PrometheusAlertRule
Type
Enter a unique, descriptive name for your Prometheus alert rule. Choose a name that clearly indicates the condition being monitored. Please note that reserved alert names for TrueFoundry managed alerts are not allowed. For checking the list of reserved alert names, please refer to the documentation.
Enter a valid PromQL expression that defines the condition for triggering this alert. The alert will fire when this expression evaluates to true for the duration specified in the duration to trigger alert field.
The prometheus expression must remain true for this duration (in seconds) before the alert is triggered. If the condition becomes false before this time elapses, the alert will not fire.
Description of the alert rule which will be displayed in the alert rule list. This can be used to provide more context about the alert rule.
When enabled, notifications will be sent to all configured target channels when the alert conditions are met.
Prompt
PromptFooGuardrailConfig
The name of the Guardrail Config.
PromptVersion
ProviderAccounts
Union, One Of:PublicCostMetric
PySparkTaskConfig
Extra configuration properties to be passed to the spark job. Docs
Configure environment variables to be injected in the task either as plain text or secrets. Docs
🔗 VolumeMountConfigure data to be mounted to Workflow pod(s) as a volume.
Service Account
PythonBuild
Python version to run your application. Should be one of the tags listed on Official Python Docker Page
Build path relative to project root path.
Path to
requirements.txt relative to Path to build contextDefine pip package requirements. In Python/YAML E.g. [“fastapi>=0.90,<1.0”, “uvicorn”]
🔗 PythonBuildPythonDependenciesPython dependencies to install
Debian packages to install via
apt get. In Python/YAML E.g. [“git”, “ffmpeg”, “htop”]🔗 PythonBuildCommandCommand to run when the container starts. Command will be set as the Entrypoint of the generated image. When deploying a Job, the command can be templatized by defining
params and referencing them in command E.g. python main.py --learning_rate \{\{learning_rate\}\}Version of CUDA Toolkit and CUDNN to install in the image These combinations are based off of publically available docker images on docker hub You can also specify a valid tag of the form {cuda_version_number}-cudnn{cudnn_version_number}-{runtime|devel}-ubuntu{ubuntu_version} Refer https://hub.docker.com/r/nvidia/cuda/tags for valid set of values Note: We use deadsnakes ubuntu ppa to add Python that currently supports only Ubuntu 18.04, 20.04 and 22.04
PythonBuildCommand
Union, One Of:PythonBuildPythonDependencies
Union, One Of:PythonTaskConfig
🔗 PythonTaskConfigImageSpecify the image spec for the task
Configure environment variables to be injected in the task either as plain text or secrets. Docs
🔗 PythonTaskConfigMountsItemConfigure data to be mounted to Workflow pod(s) as a volume.
Service Account
PythonTaskConfigImage
Union, One Of:PythonTaskConfigMountsItem
Union, One Of:PyTorchFramework
Type of the framework
QuayArtifactsRegistry
The name of the integration that will be displayed in the TrueFoundry UI.
The URL of the registry.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
QuayBasicAuth
The username for the basic authentication.
The password for the basic authentication.
QuayProviderAccount
The name of the provider account.
The name of the account on Quay.
List of integrations that are associated with the provider account.
QuerySpansResponse
🔗 TraceSpanArray of flat spans
🔗 TokenPaginationPagination information
RateLimitConfig
Name of the rate limiting configuration
🔗 RateLimitRuleList of rate limit rules
RateLimitRule
Unique identifier for the rule
Maximum number of requests or tokens allowed
RateLimitWhen
List of subjects that this rule applies to
List of models that this rule applies to
Metadata key-value pairs that this rule applies to
Recommendation
recommendation_data
typing_extensions.Annotated[(typing.Dict[(str, typing.Optional[typing.Any])], Any)]
RefusalContentPart
Type of the content part
Reason
RegisterUsersResponse
Union, One Of:RemoteSource
Remote repository URI
Resources
Requested CPU which determines the minimum cost incurred. The CPU usage can exceed the requested amount, but not the value specified in the limit. 1 CPU means 1 CPU core. Fractional CPU can be requested like
0.5 or 0.05CPU limit beyond which the usage cannot be exceeded. 1 CPU means 1 CPU core. Fractional CPU can be requested like
0.5. CPU limit should be >= cpu request.Requested memory which determines the minimum cost incurred. The unit of memory is in megabytes(MB). So 1 means 1 MB and 2000 means 2GB.
Memory limit after which the application will be killed with an OOM error. The unit of memory is in megabytes(MB). So 1 means 1 MB and 2000 means 2GB. MemoryLimit should be greater than memory request.
Requested disk storage. The unit of memory is in megabytes(MB). This is ephemeral storage and will be wiped out on pod restarts or eviction
Disk storage limit. The unit of memory is in megabytes(MB). Exceeding this limit will result in eviction. It should be greater than the request. This is ephemeral storage and will be wiped out on pod restarts or eviction
Define the shared memory requirements for your workload. Machine learning libraries like Pytorch can use Shared Memory for inter-process communication. If you use this, we will mount a
tmpfs backed volume at the /dev/shm directory. Any usage will also count against the workload’s memory limit (resources.memory_limit) along with your workload’s memory usage. If the overall usage goes above resources.memory_limit the user process may get killed. Shared Memory Size cannot be more than the defined Memory Limit for the workload.🔗 ResourcesNodeThis field determines how the underlying node resource is to be utilized
🔗 ResourcesDevicesItemDefine custom device or accelerator requirements for your workload. We currently support NVIDIA GPUs, AWS Inferentia Accelerators, Single Host TPU Slices.
ResourcesDevicesItem
Union, One Of:ResourcesNode
Union, One Of:RetryConfig
Number of attempts to retry the request
Delay between retries in milliseconds
Status codes to retry on
RevokeAllPersonalAccessTokenResponse
Union, One Of:Rolling
Percentage of total replicas that can be brought down at one time. For a value of 25 when replicas are set to 12 this would mean minimum (25% of 12) = 3 pods might be unavailable during the deployment. Setting this to a higher value can help in speeding up the deployment process.
Percentage of total replicas of updated image that can be brought up over the total replicas count. For a value of 25 when replicas are set to 12 this would mean (12+(25% of 12) = 15) pods might be running at one time. Setting this to a higher value can help in speeding up the deployment process.
RpsMetric
Average request per second averaged over all replicas that autoscaler should try to maintain
RStudio
SagemakerModel
Name to identify this Sagemaker model in the UI
The unique identifier for the Sagemaker model
🔗 ModelTypeSpecify the type of the Sagemaker model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
SambaNovaKeyAuth
The API key for SambaNova authentication
SambaNovaModel
Name to identify this SambaNova model in the UI
The name of the SambaNova model to use
🔗 ModelTypeSpecify the type of the SambaNova model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
SambaNovaProviderAccount
The name of the SambaNova provider account
List of integrations that are associated with the SambaNova provider account
🔗 CollaboratorList of users who have access to this provider account
Schedule
Specify the schedule for this job to be run periodically in cron format.
* * * * * | | | | | | | | | |___ day of week (0-6) (Sunday is 0) | | | |_____ month (1-12) | | |_______ day of month (1-31) | |_________ hour (0-23) |___________ minute (0-59)🔗 ScheduleConcurrencyPolicyChoose whether to allow this job to run while another instance of the job is running, or to replace the currently running instance. Allow will enable multiple instances of this job to run. Forbid will keep the current instance of the job running and stop a new instance from being run. Replace will terminate any currently running instance of the job and start a new one.
Timezone against which the cron schedule will be calculated, e.g. “Asia/Tokyo”. Default is machine’s local time. https://docs.truefoundry.com/docs/list-of-supported-timezones
Secret
SecretGroup
SecretInput
Key of the secret.
Value of the secret.
SecretMount
Absolute file path where the file will be created.
The TrueFoundry secret whose value will be the file content.
SecretStoreConfig
The secret store to store the virtual account token in
The path where the virtual account token will be stored in the secret store (example: secret/path/to/virtual-account-token)
SecretVersion
SelfHostedModel
2 to 62 characters long alphanumeric word, may contain - or . in between, cannot start with a number
The name of the hosted model
enter valid https/http URL that should not end with trailing slash
🔗 SelfHostedModelModelServerThe type of model server being used
🔗 ModelTypeSpecify the type of the model
🔗 SelfHostedModelAuthDataSelfHostedModel authentication data for the integration.
SelfHostedModel headers for the integration. Forwarded to the provider as is. For example:
\{"Authorization": "APIKey \<token\>"\}List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
SelfHostedModelAuthData
Union, One Of:SelfHostedModelProviderAccount
The name of the provider account.
🔗 SelfHostedModelList of integrations that are associated with the provider account.
🔗 CollaboratorCollaborators
Service
🔗 ServiceReplicasDeploy multiple instances of your pods to distribute incoming traffic across them, ensuring effective load balancing.
Whether to allow intercepts to be applied for this service. This would inject an additional sidecar in each pod of the service. Not recommended on production
🔗 ServiceRolloutStrategyStrategy to dictate how a rollout should happen when a new release for this service is made Docs
ServiceAutoscaling
🔗 ServiceAutoscalingMetricsMetrics to use for the autoscaler
ServiceAutoscalingMetrics
Union, One Of:ServiceReplicas
Union, One Of:ServiceRolloutStrategy
Union, One Of:Session
SignedUrl
SklearnFramework
Type of the framework
Relative path to the model file in the model version contents
SklearnModelSchema
🔗 InferMethodNameName of the method used for inference
Schema of the input
Schema of the output
SlackBot
Specify the notification channel to send alerts to
List of channels to send messages to.
SlackBotAuth
The bot token for the slack bot to send messages to the channels. This bot should have the following permissions: chat:write, chat:write.public
SlackBotIntegration
The name of the integration that will be displayed in the TrueFoundry UI.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
SlackIntegrations
Union, One Of:SlackProviderAccount
The name of the provider account.
🔗 SlackIntegrationsList of integrations that are associated with the provider account.
SlackWebhook
Specify the notification channel to send alerts to
SlackWebhookAuth
The URL of the slack webhook
SlackWebhookIntegration
The name of the integration that will be displayed in the TrueFoundry UI.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
SlaCutoff
Time per output token threshold in milliseconds. If time per output token exceeds this value, the target will be marked as unhealthy
SmtpCredentials
The SMTP host of the email server.
The username for the SMTP server.
The password for the SMTP server.
The port of the SMTP server.
Whether to use TLS for the SMTP server.
SpaCyFramework
Type of the framework
SparkBuild
Spark version should match the spark version installed in the image.
Build path relative to project root path.
Path to
requirements.txt relative to Path to build contextSparkConfig
The base domain for the cluster with which you can access your Spark UI
SparkDriverConfig
SparkExecutorConfig
🔗 SparkExecutorConfigInstancesExecutor Instances
SparkExecutorConfigInstances
Union, One Of:SparkExecutorDynamicScaling
Minimum number of instances to start / scale down to
Maximum number of instances to scale up to
SparkExecutorFixedInstances
Number of instances to start
SparkImage
Spark version should match the spark version installed in the image.
The image URI. Specify the name of the image and the tag. If the image is in Dockerhub, you can skip registry-url (for e.g.
tensorflow/tensorflow). You can use an image from a private registry using Advanced fieldsFQN of the container registry. If you can’t find your registry here, add it through the Integrations page
SparkImageBuild
FQN of the container registry. If you can’t find your registry here, add it through the Integrations page
🔗 SparkImageBuildBuildSourceFetch source code
SparkImageBuildBuildSource
Union, One Of:SparkJob
Name of the job
🔗 SparkJobImageThe image to use for driver and executors. Must have spark installed. Spark version must match the version in the image.
🔗 SparkJobEntrypointEntrypoint
Configure environment variables to be injected in the service either as plain text. Docs
Extra configuration properties to be passed to the spark job. Docs
🔗 VolumeMountConfigure volumes to be mounted to driver and executors. Docs
Specify the maximum number of attempts to retry a job before it is marked as failed.
Service Account
Fully qualified name of the workspace
SparkJobEntrypoint
Union, One Of:- 🔗 SparkJobPythonEntrypoint
- 🔗 SparkJobScalaEntrypoint
- 🔗 SparkJobJavaEntrypoint
- 🔗 SparkJobPythonNotebookEntrypoint
- 🔗 SparkJobScalaNotebookEntrypoint
SparkJobImage
Union, One Of:SparkJobJavaEntrypoint
The main application file to be executed by the spark job.
The main class to be executed by the spark job.
Arguments to be passed to the main application file.
SparkJobPythonEntrypoint
The main application file to be executed by the spark job. Relative path in case of git repository.
Arguments to be passed to the main application file.
SparkJobPythonNotebookEntrypoint
The main application file to be executed by the spark job. Relative path in case of git repository.
SparkJobScalaEntrypoint
The main application file to be executed by the spark job.
The main class to be executed by the spark job.
Arguments to be passed to the main application file.
SparkJobScalaNotebookEntrypoint
The main application file to be executed by the spark job. Relative path in case of git repository.
SparkJobTriggerInput
Main Class for Spark Job
Main Application File for Spark Job
Arguments to pass to the main application file
SqsInputConfig
AWS SQS Queue URL of Subscriber
AWS Region Name
A period during which Amazon SQS prevents all consumers from receiving and processing the message. If one message takes 5 seconds to process, you can set this number to 7 or any number higher than 5. This will ensure that while the message is being processed, it will not be available to other replicas. For more information, see here
Wait timeout for long polling. For more information, see here
SqsOutputConfig
AWS SQS Queue URL of Publisher
AWS Region Name
SqsQueueMetricConfig
Upper limit of the number of backlog messages the auto-scaler will try to maintain per replica. If you set this number to 10 and have 30 messages in the queue and one replica, the auto-scaler will scale the number of replicas to 3.
SshServer
Add Your SSH Public Key, this will be used to authenticate you to the SSH Server. You can find it using
cat ~/.ssh/id_rsa.pub in Mac/Linux or type $home\.ssh\id_rsa.pub in Windows Powershell. You can also generate a new SSH key pair using ssh-keygen -t rsa in your local terminal. (same for both Mac/Linux and Windows Powershell)Stop the SSH Server instance after this much time in minutes of inactivity. The instance is considered active if there is at least one active SSH connection (a client connected to the SSH server), or if a background job is running using tmux or screen, or if the pod has restarted.
SshServerConfig
The base domain for the cluster with which you can access your SSH containers
The port for the cluster with which you can access your SSH containers
SsoTeamManifest
Name of the Team
SSO configuration to use for this team
Group name from SSO that will be used to determine team membership
StageArtifactRequestManifest
Union, One Of:StageArtifactResponse
StaticVolumeConfig
Volume Type for the volume.
Persistent Volume Name of the volume to be used.
StatsModelsFramework
Type of the framework
StringDataMount
Absolute file path where the file will be created.
The file content.
Subject
Subject ID
🔗 SubjectTypeSubject type
Subject slug
Subject display name
SystemMessage
Role of the message
🔗 SystemMessageContentText content for the system message
Name of the system
SystemMessageContent
Union, One Of:TaskDockerFileBuild
FQN of the container registry. If you can’t find your registry here, add it through the Integrations page
The file path of the Dockerfile relative to project root path.
Build arguments to pass to docker build
TaskPySparkBuild
Spark version should match the spark version installed in the image.
FQN of the container registry. If you can’t find your registry here, add it through the Integrations page
Path to
requirements.txt relative to Path to build contextDefine pip package requirements. In Python/YAML E.g. [“fastapi>=0.90,<1.0”, “uvicorn”]
Debian packages to install via
apt get. In Python/YAML E.g. [“git”, “ffmpeg”, “htop”]TaskPythonBuild
FQN of the container registry. If you can’t find your registry here, add it through the Integrations page
Python version to run your application. Should be one of the tags listed on Official Python Docker Page
Path to
requirements.txt relative to Path to build contextDefine pip package requirements. In Python/YAML E.g. [“fastapi>=0.90,<1.0”, “uvicorn”]
Debian packages to install via
apt get. In Python/YAML E.g. [“git”, “ffmpeg”, “htop”]Version of CUDA Toolkit and CUDNN to install in the image These combinations are based off of publically available docker images on docker hub You can also specify a valid tag of the form {cuda_version_number}-cudnn{cudnn_version_number}-{runtime|devel}-ubuntu{ubuntu_version} Refer https://hub.docker.com/r/nvidia/cuda/tags for valid set of values Note: We use deadsnakes ubuntu ppa to add Python that currently supports only Ubuntu 18.04, 20.04 and 22.04
Team
TeamManifest
Name of the Team
Emails of each of the user who can manage the members of the team. Any tenant admin is by default a team manager.
Enter email of each of the user you want to add in the team.
TensorFlowFramework
Type of the framework
TerminateJobResponse
Terminate Job Message
🔗 JobRunStatusJob run status
TextContentPart
Type of the content part
🔗 TextContentPartTextText content for the message
TextContentPartText
Union, One Of:TogetherAiKeyAuth
Your Together AI API key for authentication
TogetherAiModel
Name to identify this Together AI model in the UI
The name of the Together AI model to use
🔗 ModelTypeSpecify the type of the Together AI model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
TogetherAiProviderAccount
The name of the Together AI provider account
List of integrations that are associated with the Together AI provider account
🔗 CollaboratorList of users who have access to this provider account
TokenPagination
Number of items per page
Base64 encoded token for the next page
Base64 encoded token for the previous page
ToolCall
Type of the tool call
Unique identifier for the tool call
ToolMessage
Role of the message
🔗 ToolMessageContentContent of the tool call result
Unique identifier for the tool call
ToolMessageContent
Union, One Of:ToolSchema
Type of the tool
TraceSpan
Unique identifier for the span within the trace.
Unique identifier for the trace that contains this span.
Identifier of the parent span in the trace hierarchy.
Name of the service that generated this span.
Name of the span
Type of span (e.g., CLIENT, SERVER, INTERNAL, PRODUCER, CONSUMER).
Name of the instrumentation scope that created this span.
Version of the instrumentation scope that created this span.
Timestamp in ISO 8601 format (e.g., 2025-03-12T00:00:09.872Z).
Duration of the span in nanoseconds.
Status code of the span (e.g., OK, ERROR, UNSET).
Human-readable status message describing the span result.
Key-value pairs containing additional metadata about the span.
Array of events that occurred during the span execution.
🔗 SubjectSubject (user or virtualaccount) that created this span.
TracingProject
TracingProjectManifest
Name of the tracing project
Name of the ML Repo
TransformersFramework
Type of the framework
🔗 LibraryNameName of the library for the framework
The
pipeline() task this model can be used with e.g. text-generation. See huggingface docs for all possible valuesBase model Id from HuggingFace Hub. If this is a finetuned model, this points to the base model id used for finetuning.
TriggerJobRequestInput
Union, One Of:TriggerJobRunResponse
Job triggered
Name of the job run
TrueFoundryApplyRequestManifest
Union, One Of:- 🔗 MlRepoManifest
- 🔗 ArtifactManifest
- 🔗 ModelManifest
- 🔗 ChatPromptManifest
- 🔗 DataDirectory
- 🔗 Service
- 🔗 ApplicationSet
- 🔗 ProviderAccounts
- 🔗 ClusterManifest
- 🔗 WorkspaceManifest
- 🔗 Job
- 🔗 Helm
- 🔗 Volume
- 🔗 Notebook
- 🔗 RStudio
- 🔗 Workflow
- 🔗 AsyncService
- 🔗 SshServer
- 🔗 SparkJob
- 🔗 GatewayConfig
- 🔗 TeamManifest
- 🔗 PolicyManifest
- 🔗 AlertConfig
- 🔗 VirtualAccountManifest
- 🔗 CommonToolsSettings
- 🔗 AiFeaturesSettings
TrueFoundryApplyResponse
existing_manifest
typing_extensions.Annotated[(typing.Optional[TrueFoundryApplyResponseExistingManifest], Any)]
🔗 TrueFoundryApplyResponseExistingManifestThe existing manifest of the resource
🔗 TrueFoundryApplyResponseActionThe action performed: CREATE or UPDATE
The created or updated resource
TrueFoundryApplyResponseExistingManifest
Union, One Of:- 🔗 MlRepoManifest
- 🔗 ArtifactManifest
- 🔗 ModelManifest
- 🔗 ChatPromptManifest
- 🔗 DataDirectory
- 🔗 Service
- 🔗 ApplicationSet
- 🔗 ProviderAccounts
- 🔗 ClusterManifest
- 🔗 WorkspaceManifest
- 🔗 Job
- 🔗 Helm
- 🔗 Volume
- 🔗 Notebook
- 🔗 RStudio
- 🔗 Workflow
- 🔗 AsyncService
- 🔗 SshServer
- 🔗 SparkJob
- 🔗 GatewayConfig
- 🔗 TeamManifest
- 🔗 PolicyManifest
- 🔗 AlertConfig
- 🔗 VirtualAccountManifest
- 🔗 CommonToolsSettings
- 🔗 AiFeaturesSettings
TrueFoundryArtifactSource
Artifact or Model Version FQN of the artifact to be downloaded
Environment variable which will contain the download path of the artifact
TrueFoundryDbssm
The name of the integration that will be displayed in the TrueFoundry UI.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
TrueFoundryDeleteRequestManifest
Union, One Of:- 🔗 MlRepoManifest
- 🔗 ArtifactManifest
- 🔗 ModelManifest
- 🔗 ChatPromptManifest
- 🔗 DataDirectory
- 🔗 Service
- 🔗 ApplicationSet
- 🔗 ProviderAccounts
- 🔗 ClusterManifest
- 🔗 WorkspaceManifest
- 🔗 Job
- 🔗 Helm
- 🔗 Volume
- 🔗 Notebook
- 🔗 RStudio
- 🔗 Workflow
- 🔗 AsyncService
- 🔗 SshServer
- 🔗 SparkJob
- 🔗 GatewayConfig
- 🔗 TeamManifest
- 🔗 PolicyManifest
- 🔗 AlertConfig
- 🔗 VirtualAccountManifest
- 🔗 CommonToolsSettings
- 🔗 AiFeaturesSettings
TrueFoundryInteractiveLogin
List of paths that will bypass auth. needs to start with a forward slash(/) and should not contain wildcards(*)
TrueFoundryManagedSource
Type of the source
TrueFoundryProviderAccount
The name of the provider account.
List of integrations that are associated with the provider account.
TtlProviderAccount
The name of the provider account.
List of integrations that are associated with the provider account.
TtlRegistry
The name of the integration that will be displayed in the TrueFoundry UI.
The URL of the registry.
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
UpdateSecretInput
Key of the secret.
Value of the secret.
UpdateUserRolesResponse
Union, One Of:UpgradeData
UsageCodeSnippet
Libraries used in the code snippet
User
UserMessage
Role of the message
🔗 UserMessageContentContent of the user message. can be a mix of text and images
Name of the user this message is from
UserMessageContent
Union, One Of:UserMessageContentItem
Union, One Of:UserMetadata
user_object
typing_extensions.Annotated[(typing.Optional[typing.Dict[(str, typing.Optional[typing.Any])]], Any)]
tenant_role_managed_by
typing_extensions.Annotated[(typing.Optional[UserMetadataTenantRoleManagedBy], Any)]
UserResource
Resource Type
Resource ID
Role ID
Resource Name (if available)
Uv
UV version to use
Sync options to pass to uv command
ValidationError
ValidationErrorLocItem
Union, One Of:VertexModel
Display Name - 2 to 62 characters long alphanumeric word, may contain - or . in between, cannot start with a number
🔗 ModelTypeSpecify the type of the model
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
VertexModelV2
Name to identify this Vertex AI model in the UI
The unique identifier for the Vertex AI model
🔗 ModelTypeSpecify the type of the Vertex AI model (e.g., chat, text, etc.)
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
VirtualAccount
VirtualAccountManifest
Virtual Account Name
Expiration Date of the Virtual Account (should be in the format yyyy-mm-dd)
🔗 PermissionsPermissions
VirtualMcpServerIntegration
The name of the Virtual MCP Server.
Provide a brief description of the purpose of this Virtual MCP Server.
🔗 VirtualMcpServerSourceList of source MCP servers to include in this virtual server.
List of teams/users who can use this MCP server.
VirtualMcpServerSource
Name of the source MCP server within this group (use {provider_account_name}/{integration_name} format).
List of tools to enable from this source server. Leave empty to enable all tools.
Volume
Name of the Volume. This will be set as the volume name.
🔗 VolumeConfigVolume Config - Volume Configuration, can be either Dynamically provisioned or statically provisioned.
Fully qualified name of the workspace
VolumeBrowser
Username for logging in the volume browser.
TFY Secret containing the password for logging in the volume browser.
Kubernetes Service account name for the volume browser.
VolumeConfig
Union, One Of:VolumeMount
Absolute file path where the volume will be mounted.
Sub path within the volume to mount. Defaults to root of the volume.
The TrueFoundry volume that needs to be mounted.
WebhookBasicAuth
Basic auth username
Basic auth password
WebhookBearerAuth
Bearer token for authentication
Prefix for the token
WebhookIntegration
The name of the integration that will be displayed in the TrueFoundry UI.
The URL of the webhook endpoint
🔗 WebhookIntegrationAuthDataAuthentication configuration for the webhook
List of subjects that are authorized to access this integration. List of user fqn in format <user_type>:<username>.
WebhookIntegrationAuthData
Union, One Of:WebhookProviderAccount
The name of the provider account.
🔗 WebhookIntegrationList of integrations that are associated with the provider account.
WeightBasedLoadBalancing
Routing Type
🔗 LoadBalanceTargetList of targets for load balancing with weights
WeightBasedLoadBalancingRule
Unique identifier for the rule
WorkbenchImage
The image URI. Specify the name of the image and the tag. If the image is in Dockerhub, you can skip registry-url (for e.g.
tensorflow/tensorflow). You can use an image from a private registry using Advanced fieldsThe build script to run when building the image. This will be executed as the last step in the docker build process as the root user (RUN DEBIAN_FRONTEND=noninteractive bash -ex build_script.sh)
FQN of the container registry. If you can’t find your registry here, add it through the Integrations page
WorkerConfig
🔗 WorkerConfigInputConfigInput Config
🔗 WorkerConfigOutputConfigOutput Config
Number of concurrent workers to spawn for the processor
WorkerConfigInputConfig
Union, One Of:WorkerConfigOutputConfig
Union, One Of:Workflow
Name of the workflow
🔗 WorkflowSourceSource Code for the workflow, either local or remote
Path to the workflow file relative to the project root path
🔗 WorkflowFlyteEntitiesItemFlyte Entities
🔗 WorkflowAlertAlerts
WorkflowAlert
Send an alert when the job completes
Send an alert when the job fails
WorkflowFlyteEntitiesItem
Union, One Of:WorkflowSource
Union, One Of:Workspace
WorkspaceManifest
Cluster FQN
Name - 3 to 32 lower case characters long alphanumeric word, may contain - in between, cannot start with a number
Tag the environment like dev, staging or production. You will need to create environments if you don’t have already.
🔗 CollaboratorCollaborators
🔗 PermissionsPermissions
WorkspacesDeleteResponse
Destroyed the workspace
XgBoostFramework
Type of the framework
Relative path to the model file in the model version contents
XgBoostModelSchema
Name of the method used for inference
Schema of the input
Schema of the output