Type Alias: ProviderBatchEvent

ProviderBatchEvent = { batch: ProviderBatchReference; modelId: string; requestIds: string[]; type: "submitted"; } | { batch: ProviderBatchReference; requestCounts?: BatchRequestCounts; status: "pending" | "completed" | "failed"; type: "status"; } | { batch: ProviderBatchReference; type: "completed"; } | { batch?: ProviderBatchReference; error: unknown; requestIds: string[]; type: "failed"; } | { batch: ProviderBatchReference; type: "cancelled"; } | { batch: ProviderBatchReference; type: "resumed"; } | { modelId: string; provider: string; reason: string; requestIds?: string[]; type: "fallback"; }

Defined in: packages/core/src/adapters/ai-sdk/provider-batch.ts:133

Lifecycle of the provider batches a batch({ providerBatch }) run submits.

Union Members

Type Literal

{ batch: ProviderBatchReference; modelId: string; requestIds: string[]; type: "submitted"; }

---

Type Literal

{ batch: ProviderBatchReference; requestCounts?: BatchRequestCounts; status: "pending" | "completed" | "failed"; type: "status"; }

---

Type Literal

{ batch: ProviderBatchReference; type: "completed"; }

---

Type Literal

{ batch?: ProviderBatchReference; error: unknown; requestIds: string[]; type: "failed"; }

---

Type Literal

{ batch: ProviderBatchReference; type: "cancelled"; }

---

Type Literal

{ batch: ProviderBatchReference; type: "resumed"; }

A replayed durable run found calls it had already submitted, and is reading their answers from this batch rather than submitting again.

---

Type Literal

{ modelId: string; provider: string; reason: string; requestIds?: string[]; type: "fallback"; }

Calls went out live at the standard price: the model's provider has no batch endpoint, or refused this submission as unsupported (requestIds is then set).