Inference cluster management.
A cluster groups multiple registered hosts to serve an LLM model via an
OpenAI-compatible endpoint at /inference/{slug}/v1/chat/completions.
Summary
Functions
Create an inference cluster.
Delete a cluster.
Get a specific cluster.
List inference clusters.
Start a stopped cluster.
Stop a running cluster.
Types
@type result() :: {:ok, map()} | {:error, Miosa.Error.t()}
Functions
@spec create(Miosa.Client.t(), map()) :: result()
Create an inference cluster.
attrs must include :name, :model, :host_ids.
@spec delete(Miosa.Client.t(), String.t()) :: result()
Delete a cluster.
@spec get(Miosa.Client.t(), String.t()) :: result()
Get a specific cluster.
@spec list(Miosa.Client.t()) :: result()
List inference clusters.
@spec start(Miosa.Client.t(), String.t()) :: result()
Start a stopped cluster.
@spec stop(Miosa.Client.t(), String.t()) :: result()
Stop a running cluster.