TransferTimeoutPolicy

Defined in: src/transfers/TransferJob.ts:74

Timeout policy applied by the transfer engine.

Two timeout scopes exist with deliberately different failure semantics:

• Job scope (timeoutMs): covers the full engine execution including retries. When it fires, the engine rethrows the TimeoutError immediately - the retry policy is never consulted.
• Attempt scope (attemptTimeoutMs and stallTimeoutMs): covers a single attempt. When either fires, the per-attempt abort controller cancels the attempt and the resulting TimeoutError flows into the retry policy like any other attempt failure, so retryable timeouts are retried (with backoff) instead of failing the job.

Example

await engine.execute(job, executor, {
  retry: createDefaultRetryPolicy(),
  timeout: { timeoutMs: 600_000, attemptTimeoutMs: 120_000, stallTimeoutMs: 30_000 },
});

Properties

Property	Type	Description	Defined in
attemptTimeoutMs?	number	Maximum duration for a single attempt in milliseconds. Expiry aborts only the active attempt; the failure flows into the retry policy.	src/transfers/TransferJob.ts:81
retryable?	boolean	Whether timeout failures are retryable. Defaults to true.	src/transfers/TransferJob.ts:89
stallTimeoutMs?	number	Maximum time without progress before an attempt is considered stalled, in milliseconds. The watchdog resets on every progress report; expiry aborts only the active attempt and the failure flows into the retry policy.	src/transfers/TransferJob.ts:87
timeoutMs?	number	Maximum duration for the full engine execution, including retries, in milliseconds.	src/transfers/TransferJob.ts:76