Skip to main content

vantage_sdk.workbench.dynamo_deployment

SDK helpers for cluster-scoped NVIDIA Dynamo deployment operations.

Classes

DynamoDeploymentSDK()

SDK for cluster-scoped Dynamo platform + deployment operations.

Methods

  • create(self, ctx, *, cluster_name: 'str', name: 'str', model: 'str', namespace: 'str | None' = None, backend: 'str' = 'auto', search_strategy: 'str' = 'rapid', auto_apply: 'bool' = True, ttft: 'float | None' = None, itl: 'float | None' = None, isl: 'int | None' = None, osl: 'int | None' = None, request_rate: 'float | None' = None, concurrency: 'int | None' = None, configuration_preset: 'str | None' = None, compute_pool: 'str | None' = None, frontend_compute_pool: 'str | None' = None, router_mode: 'str | None' = None, planner: 'dict[str, Any] | None' = None, dgd_overrides: 'dict[str, Any] | None' = None) -> 'ClusterServiceResponse': Create a Dynamo model deployment (renders a DGDR CR).
  • delete(self, ctx, *, cluster_name: 'str', name: 'str') -> 'ClusterServiceResponse': Delete a Dynamo deployment (DGDR + generated DGD).
  • get(self, ctx, *, cluster_name: 'str', name: 'str') -> 'ClusterServiceResponse': Get one Dynamo deployment by name.
  • list(self, ctx, *, cluster_name: 'str') -> 'ClusterServiceResponse': List the caller's Dynamo deployments.
  • status(self, ctx, *, cluster_name: 'str') -> 'ClusterServiceResponse': Get Dynamo platform status (deployed bool + component detail).