Node.js "AbortError: The operation was aborted" - Fix Timeout Aborts in CI

An operation tied to an AbortSignal was cancelled - usually a timeout (AbortSignal.timeout) firing because a fetch or async task took longer than its budget. In CI this is often a slow network rather than a stuck operation.

What this error means

A fetch or other abortable call rejects with AbortError: The operation was aborted (code ABORT_ERR/ERR_ABORTED). It can be intermittent - passing on retry - when the cause is a transiently slow request hitting a tight timeout.

DOMException [AbortError]: The operation was aborted
    at node:internal/deps/undici/undici:...
  code: 20,
  name: 'AbortError'

Common causes

A timeout fired on a slow but legitimate request

A short AbortSignal.timeout(ms) cancels a fetch that was simply slow on a congested CI network. The operation would have succeeded with more time.

The signal was aborted by app logic

Code explicitly called controller.abort() (e.g. on shutdown or a superseded request) and a pending operation tied to that signal rejected as designed.

How to fix it

Raise the timeout and retry transient aborts

Give slow CI networks more headroom and retry timeout-driven aborts.

async function get(url) {
  for (let i = 1; i <= 3; i++) {
    try {
      return await fetch(url, { signal: AbortSignal.timeout(15000) });
    } catch (err) {
      if (err.name !== 'AbortError' || i === 3) throw err;
    }
  }
}

Distinguish timeout from intentional abort

1. Check whether the abort came from your own controller.abort() or from a timeout signal.
2. For intentional aborts (shutdown, superseded request), treat the rejection as expected, not an error.
3. For timeouts, raise the budget or retry rather than failing the job.

How to prevent it

• Size AbortSignal.timeout budgets for the slowest realistic CI network.
• Retry timeout-driven aborts; do not retry intentional aborts.
• Cache or mock external calls in tests so they do not depend on live latency.

Related guides