Puppeteer: net::ERR_ABORTED at <url>

Chrome started the navigation and then cancelled it before a document was committed. page.goto() reports that as net::ERR_ABORTED. Unlike DNS or connection errors, the server may have responded perfectly; the browser simply decided not to display the response as a page.

The best-known trigger is navigating directly to a file that Chrome downloads instead of rendering, such as a PDF, ZIP or CSV served with Content-Disposition: attachment. In headless mode Chrome does not render PDFs as a page, so goto on a PDF URL aborts. Another trigger is a second navigation (a redirect in JavaScript, another goto, or a click) that replaces the first before it finishes. Request interception that calls request.abort() on the document request also produces it.

What is not wrong: the URL usually works in a normal browser, where it would start a download.

// Download a file using the page's session instead of navigating to it
const data = await page.evaluate(async url => {
  const res = await fetch(url, { credentials: 'include' });
  const buf = await res.arrayBuffer();
  return Array.from(new Uint8Array(buf));
}, 'https://example.com/report.pdf');
await fs.promises.writeFile('report.pdf', Buffer.from(data));

Treat file URLs as downloads, not pages. Keep a single navigation in flight per page, and keep interception rules scoped to subresources. When crawling unknown links, check the content type with a HEAD request before navigating, and route file types such as PDF and CSV to a download path instead of page.goto().