NextJS 14 returns fetch failed with UND_ERR_CONNECT_TIMEOUT error on serverless function

[andremendonca03]

Link to the code that reproduces this issue

https://codesandbox.io/p/devbox/gifted-shirley-mzlvgy?file=%2Fapp%2Fapi%2Froute.js%3A1%2C1-39%2C1

To Reproduce

1. From a client-side component, start a fecth POST request to an API endpoint (route handler) on form submission;
2. On an API serverless function realise another fetch POST request to an external API (in my case I used Slack message API);
3. On a production environment hosted on Vercel, around 70% of the requests to Slack are working while another 30% fail returning 500 server error code "UND_ERR_CONNECT_TIMEOUT".

Systems:

• Next.js v14.1.4
• Errors happen in both Pages and App Router
• Could only reproduce on a Prod environment hosted on Vercel
• Serverless function running node v20x
• Requests to external APIs from a client component work fine. Only serverless functions error out
• Slack API is working fine. Tested on other pure React apps and Postman and had no errors at all
• Problem seems to happen only on external API calls (tested Mailchimp as well and failed less times than Slack but still failing);

Full error message:
Unhandled Rejection: TypeError: fetch failed at node:internal/deps/undici/undici:12345:11 at process.processTicksAndRejections (node:internal/process/task_queues:95:5) { cause: ConnectTimeoutError: Connect Timeout Error at onConnectTimeout (node:internal/deps/undici/undici:7492:28) at node:internal/deps/undici/undici:7448:50 at Immediate._onImmediate (node:internal/deps/undici/undici:7480:13) at process.processImmediate (node:internal/timers:478:21) at process.callbackTrampoline (node:internal/async_hooks:130:17) { code: 'UND_ERR_CONNECT_TIMEOUT' } } Node.js process exited with exit status: 128. The logs above can help with debugging the issue.

Current vs. Expected behavior

Currently some external API calls from a serverless function are returning unhandled fetch errors.
Expected behaviour is no errors being returned and API call succeeding.

Provide environment information

Operating System:
Vercel Servers

Binaries:
Node: v20x (default vercel v20 setting)
npm: 10.2.3
yarn: 1.22.19
build command: yarn build

Relevant Packages: 
next: 14.1.4
react: 18.2.0
react-dom: 18.2.0

Next.js Config:
/** @type {import('next').NextConfig} */
const nextConfig = {
  reactStrictMode: true,
  trailingSlash: true,
  images: {
    remotePatterns: [
      {
        protocol: 'https',
        hostname: 'site.com',
        port: '',
        pathname: '/wp-content/uploads/**',
      },
    ],
  },
}

module.exports = nextConfig;

Which area(s) are affected? (Select all that apply)

Module Resolution, Pages Router, Runtime

Which stage(s) are affected? (Select all that apply)

Vercel (Deployed)

Additional context

Additional information about the issue and more cases can be found at this discussion:
#57384

[icyJoseph]

This is most likely an Undici error.

• https://undici.nodejs.org/#/?id=network-address-family-autoselection

Can you try to collect more data about the endpoints that are failing?

[andremendonca03]

Hey @icyJoseph I only have these 2 APIs mentioned to test at the moment.
Haven't Undici been removed from next14?

[icyJoseph]

Hi, well, it's not a Next.js thing, rather Node.js' fetch implementation uses undici at its core.

For a world pre- Node 18 (17 really) Next js did a polyfill with node-fetch, to provide server side fetch, but since Node.js adopted fetch natively, Next.js just doesn't have to anymore.

[icyJoseph]

Maybe you can do an experiment. Does it also fail, if you create a node script, or just open the Node repl, and try to make a fetch request from there?

[andremendonca03]

I'm pretty sure this issue doesn't happen on local environments, only on live servers but I'll be testing on pure node soon.

I also found these 2 related issues?
nodejs/undici#2362
nodejs/undici#1531

[hrc7505]

@andremendonca03 See #57384 (comment)

[andremendonca03]

Yeah I'm on that discussion as well. For me though it only happens in production, not local.
Which API are you trying to fetch? Is it through a client component or a serverless function? @hrc7505

[hrc7505]

@andremendonca03

1. We are getting these issues in server-rendered pages. We are fetching apis and creating pages during build.
2. Same apis are working in browser during runtime.
3. I am randomly getting this issue; not every time. Once issue occurs, for rest of the day, it behaves same. But next day again it starts working.

Node: 18.17.1
Nextjs: 14.2.3

[osama554]

Facing same issue in my project using nextjs 14.

[Lersson]

Facing the same problem with Nextjs 14.2.3 and node 20 under comporate http_proxy network. In build time, Nextjs is trying to fetch some cloudflare addresses

[cause]: ConnectTimeoutError: Connect Timeout Error (attempted addresses: 104.16.24.34:443, 104.16.28.34:443, 104.16.26.34:443, 104.16.2.35:443, 104.16.0.35:443, 104.16.27.34:443, 104.16.31.34:443, 104.16.1.35:443, 104.16.30.34:443, 104.16.25.34:443, 104.16.29.34:443, 104.16.3.35:443)

[vinc01100101]

Same problem here. All day yesterday, we were getting UND_ERR_CONNECT_TIMEOUT errors only on Vercel's production build attempts. It was working when built locally. We're still on Next13.

However, at 4:30 am, I tried to build it again in production, and everything worked fine.

This is weird. We may still encounter this issue in future builds. Hoping for someone to find a fix for this. I'll also continue to observe.

Err log:

ERROR FETCH ITEMPAGE GETSTATICPATHS: TypeError: fetch failed

at Object.fetch (node:internal/deps/undici/undici:11731:11)
at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
at async getStaticPaths (/vercel/path0/.next/server/pages/used-cars/[slug]/[itemPage].js:1471:26)
at async buildStaticPaths (/vercel/path0/node_modules/next/dist/build/utils.js:598:33)
at async /vercel/path0/node_modules/next/dist/build/utils.js:933:115
at async Span.traceAsyncFn (/vercel/path0/node_modules/next/dist/trace/trace.js:79:20) {
cause: ConnectTimeoutError: Connect Timeout Error
at onConnectTimeout (node:internal/deps/undici/undici:6869:28)
at node:internal/deps/undici/undici:6825:50
at Immediate._onImmediate (node:internal/deps/undici/undici:6857:13)
at process.processImmediate (node:internal/timers:476:21) {
code: 'UND_ERR_CONNECT_TIMEOUT'
}
}

[vinc01100101]

Alright, this suggestion nodejs/undici#1531 (comment) might be the fix for UND_ERR_CONNECT_TIMEOUT error. After some additional research and debugging, I tried using dns.lookup(), dns.resolve4(), and dns.resolve6() to determine which family (IPv4 or IPv6) the DNS I am accessing supports. It turns out that it supports IPv4.

[ornakash]

Alright, this suggestion nodejs/undici#1531 (comment) might be the fix for UND_ERR_CONNECT_TIMEOUT error. After some additional research and debugging, I tried using dns.lookup(), dns.resolve4(), and dns.resolve6() to determine which family (IPv4 or IPv6) the DNS I am accessing supports. It turns out that it supports IPv4.

Thanks. So what can I do in order to eliminate this issue?

[vinc01100101]

Thanks. So what can I do in order to eliminate this issue?

@ornakash, I just included NODE_OPTIONS=--dns-result-order=ipv4first in Vercel's environment variables for our project, and everything works fine.

Note that setting --dns-result-order=ipv4first prioritizes IPv4 addresses over IPv6 addresses but does not disregard IPv6 addresses entirely. Both IPv4 and IPv6 addresses are still resolved and included in the results, but IPv4 addresses appear first in the list. This configuration can be useful when you prefer IPv4 connectivity but still want to support IPv6.

[ornakash]

Thanks. So what can I do in order to eliminate this issue?

@ornakash, I just included NODE_OPTIONS=--dns-result-order=ipv4first in Vercel's environment variables for our project, and everything works fine.

Note that setting --dns-result-order=ipv4first prioritizes IPv4 addresses over IPv6 addresses but does not disregard IPv6 addresses entirely. Both IPv4 and IPv6 addresses are still resolved and included in the results, but IPv4 addresses appear first in the list. This configuration can be useful when you prefer IPv4 connectivity but still want to support IPv6.

Thanks! it looks like it drastically reduced the times this happens.
We had like 200 errors a day, and now only 6 with UND_ERR_CONNECT_TIMEOUT

I hope they'll understand why it happens so we won't get even 6 errors

[vinc01100101]

Thanks! it looks like it drastically reduced the times this happens. We had like 200 errors a day, and now only 6
with UND_ERR_CONNECT_TIMEOUT

I hope they'll understand why it happens so we won't get even 6 errors

@ornakash

To see which connectivity the DNS is using for your request, try using:

dns.lookup('example.com', { all: true }, (err, addresses) => {
    console.log({ addresses, err });
});

(Replace example.com with the DNS where you received the 6 errors.)

The logs should show something like:

{
  addresses: [
    { address: '93.184.215.14', family: 4 },
    { address: '2606:2800:21f:cb07:6820:80da:af6b:8b2c', family: 6 }
  ],
  err: null
}

[kpratik2015]

Linking my resolution here if it helps someone else vercel/vercel#11692 (comment)

[rafalzawadzki]

Same started happening in our project; 5% of requests on Vercel fail due to timeout. Locally they error out with UND_ERR_HEADERS_TIMEOUT after a few minutes. Affects two POST API route handlers that call external services.

Tried setting NODE_OPTIONS=--dns-result-order=ipv4first but didn't seem to help.

Running Next 14.0.4 and Node 18.x

[ngroenewold95]

I am also getting a bunch of undici errors as well recently. These are the 3 main ones

"next": "^14.2.3",
node v20.9.0

I also tried setting vercel env variable: NODE_OPTIONS=--dns-result-order=ipv4first but it has not solved the issue

`TypeError: fetch failed
    at node:internal/deps/undici/undici:12618:11
    at process.processTicksAndRejections (node:internal/process/task_queues:95:5) {
  cause: Error: connect ETIMEDOUT 76.76.21.241:443
      at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1555:16)
      at TCPConnectWrap.callbackTrampoline (node:internal/async_hooks:128:17) {
    errno: -110,
    code: 'ETIMEDOUT',
    syscall: 'connect',
    address: '76.76.21.241',
    port: 443
  }`
  

  `TypeError: fetch failed
    at node:internal/deps/undici/undici:12618:11
    at process.processTicksAndRejections (node:internal/process/task_queues:95:5) {
  cause: ConnectTimeoutError: Connect Timeout Error
      at onConnectTimeout (node:internal/deps/undici/undici:7760:28)
      at node:internal/deps/undici/undici:7716:50
      at Immediate._onImmediate (node:internal/deps/undici/undici:7748:13)
      at process.processImmediate (node:internal/timers:476:21)
      at process.callbackTrampoline (node:internal/async_hooks:128:17) {
    code: 'UND_ERR_CONNECT_TIMEOUT'
  }
}`

`TypeError: fetch failed
    at node:internal/deps/undici/undici:12618:11
    at process.processTicksAndRejections (node:internal/process/task_queues:95:5) {
  cause: [Error: C0AFB780CE7F0000:error:0A00010B:SSL routines:ssl3_get_record:wrong version number:ssl/record/ssl3_record.c:355:
  ] {
    library: 'SSL routines',
    reason: 'wrong version number',
    code: 'ERR_SSL_WRONG_VERSION_NUMBER'
  }
}`

[bacqueyrisses]

I successfully resolved the issue by configuring the undici global dispatcher in the root layout.

[ornakash]

I successfully resolved the issue by configuring the undici global dispatcher in the root layout.

Do you have a solution for pages router as well?

[ljj0915]

I successfully resolved the issue by configuring the undici global dispatcher in the root layout.

Hi ～，After I add this configuration, build will report an error
next：13.2.3
node：20.11.1

./node_modules/undici/lib/web/fetch/util.js
Module parse failed: Unexpected token (682:63)
File was processed with these loaders:

• ./node_modules/next/dist/build/webpack/loaders/next-flight-loader/index.js
• ./node_modules/next/dist/build/webpack/loaders/next-swc-loader.js
  You may need an additional loader to handle the result of these loaders.
  | // 5. If object is not a default iterator object for interface,
  | // then throw a TypeError.

        if (typeof this !== "object" || this === null || !(#target in this)) {

| throw new TypeError('next' called on an object that does not implement interface ${name} Iterator.);
| }

Import trace for requested module:
./node_modules/undici/lib/web/fetch/util.js
./node_modules/undici/lib/web/fetch/formdata.js
./node_modules/undici/index.js
./app/layout.tsx

Build failed because of webpack errors

[starlight-akouri]

I am also facing this issue when calling API functions from Zapier:

Unhandled Rejection: TypeError: fetch failed
at node:internal/deps/undici/undici:12502:13
at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
at async p (/var/task/.next/server/app/api/calls/route.js:1:634) {
[cause]: ConnectTimeoutError: Connect Timeout Error (attempted addresses: 54.203.40.250:443)
at onConnectTimeout (node:internal/deps/undici/undici:6635:28)
at node:internal/deps/undici/undici:6587:50
at Immediate._onImmediate (node:internal/deps/undici/undici:6619:13)
at process.processImmediate (node:internal/timers:478:21)
at process.callbackTrampoline (node:internal/async_hooks:130:17) {
code: 'UND_ERR_CONNECT_TIMEOUT'
}
}
Node.js process exited with exit status: 128. The logs above can help with debugging the issue.

[abhishekchoure]

I am also facing the same issue when trying to execute SQL using libsql client (Turso):

There has been an error while retrieving the database type.
Debug information:

• URL:
• Response Status Code: N/A
  ⨯ TypeError: fetch failed
  at /usr/lib/node_modules_20/undici/index-fetch.js:8:13
  at async $$ACTION_1 (./app/actions/users.ts:48:23)
  digest: "3259589146"
  Cause: AggregateError [ETIMEDOUT]:
  at internalConnectMultiple (node:net:1116:18)
  at internalConnectMultiple (node:net:1184:5)
  at Timeout.internalConnectMultipleTimeout (node:net:1710:5)
  at listOnTimeout (node:internal/timers:575:11)
  at process.processTimers (node:internal/timers:514:7) {
  code: 'ETIMEDOUT',
  [errors]: [
  Error: connect ETIMEDOUT 66.51.123.211:443
  at createConnectionError (node:net:1646:14)
  at Timeout.internalConnectMultipleTimeout (node:net:1705:38)
  at listOnTimeout (node:internal/timers:575:11)
  at process.processTimers (node:internal/timers:514:7) {
  errno: -110,
  code: 'ETIMEDOUT',
  syscall: 'connect',
  address: '66.51.123.211',
  port: 443
  },
  Error: connect ENETUNREACH 2a09:8280:1::6:d39d:443 - Local (:::0)
  at internalConnectMultiple (node:net:1180:16)
  at Timeout.internalConnectMultipleTimeout (node:net:1710:5)
  at listOnTimeout (node:internal/timers:575:11)
  at process.processTimers (node:internal/timers:514:7) {
  errno: -101,
  code: 'ENETUNREACH',
  syscall: 'connect',
  address: '2a09:8280:1::6:d39d',
  port: 443
  }
  ]
  }

[gnomefin]

Just give my data point as well, but in this isn't a serverless, just a regular next workload.

It works well when I was using these versions:

"next": "14.1.3",
 "@types/node": "^20.8.4",

But now I am. using next 14.2.4 with node ^20.8.4, this next version doesn't seem indicate to have a backward compatible, or match with some deps like undici

Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:  ⨯ TypeError: fetch failed
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:     at Object.fetch (node:internal/deps/undici/undici:11576:11)
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:     at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:     at async fetchExternalImage (/opt/commeasy/oc-web-ui/deploy/1.2.10/node_modules/next/dist/server/image-optimizer.js:565:17)
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:     at async NextNodeServer.imageOptimizer (/opt/commeasy/oc-web-ui/deploy/1.2.10/node_modules/next/dist/server/next-server.js:650:48)
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:     at async cacheEntry.imageResponseCache.get.incrementalCache (/opt/commeasy/oc-web-ui/deploy/1.2.10/node_modules/next/dist/server/next-server.js:182:65)
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:     at async /opt/commeasy/oc-web-ui/deploy/1.2.10/node_modules/next/dist/server/response-cache/index.js:90:36
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:     at async /opt/commeasy/oc-web-ui/deploy/1.2.10/node_modules/next/dist/lib/batcher.js:45:32 {
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:   cause: ConnectTimeoutError: Connect Timeout Error
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:       at onConnectTimeout (node:internal/deps/undici/undici:8522:28)
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:       at node:internal/deps/undici/undici:8480:50
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:       at Immediate._onImmediate (node:internal/deps/undici/undici:8511:13)
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:       at process.processImmediate (node:internal/timers:476:21)
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:       at process.callbackTrampoline (node:internal/async_hooks:130:17) {
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:     code: 'UND_ERR_CONNECT_TIMEOUT'
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]:   }
Jul 10 05:07:56 ip-30-0-103-10 sh[784691]: }
(END)

[ztanner]

Hi folks - just want to crosspost a response from a message shared from someone on the Vercel infra team: vercel/vercel#11692 (comment)

If you're receiving these Undici timeout errors and you're deploying with Vercel, you'll find that information helpful. If you experience it with other providers, you will likely need to wait Undici is updated in Node to include the version with the fixes.

Here's the corresponding issue on the Undici side: nodejs/undici#3410

I'll be closing this as it's not currently actionable by the Next.js team.

[github-actions]

This closed issue has been automatically locked because it had no new activity for 2 weeks. If you are running into a similar issue, please create a new issue with the steps to reproduce. Thank you.