Random HTTP 500 errors with Node process crashes (SIGABRT / SIGSEGV)

Hi everyone,

We’ve been experiencing random HTTP 500 errors in two of our production Next.js applications deployed on Vercel. The issue started around October 26th, 2025, and we haven’t been able to determine a clear cause yet.

Setup overview

App 1:

Next.js 14.2.5
Node.js 20
Fluid Compute: Disabled
Region: eu-central-1
Stable in production for several months, with no recent code changes. The issue started appearing recently without any configuration changes.

App 2:

Next.js 15.5.2
Node.js 22
Fluid Compute: Enabled
Region: eu-central-1
Newer app in production. We initially suspected Inngest, since the problem appeared soon after adding it, but the errors persist even after uninstalling it.

Both applications use the Inngest SDK, and both connect to Supabase (also hosted in the same region). Supabase logs show no connection errors or abrupt disconnections, so the database layer doesn’t appear to be involved.

Error details

The issue occurs randomly during HTTP requests. In many cases, it affects the same endpoint (the one typically hit first when users open the app), but it has also occurred on other routes.

In the Vercel logs, we consistently see entries like:

Node.js process exited with signal: 6 (SIGABRT) (core dumped)

or occasionally:

Node.js process exited with signal: 11 (SIGSEGV) (core dumped)

Immediately after, we see low-level memory errors such as:

double free or corruption (out)
corrupted double-linked list
munmap_chunk(): invalid pointer
free(): invalid pointer

Additional notes

No corresponding errors appear in our application logs.
The issue occurs sporadically, typically under normal production load (not under heavy stress).
The same code runs fine locally without crashes.
This started recently without code changes, which makes us suspect a potential platform or runtime-level issue (Node or infrastructure related).

Has anyone else seen similar behavior recently, especially in the eu-central-1 region? Any guidance or insights from would be greatly appreciated. In particular, how to debug these kind of low level errors in production

Thanks in advance

I am getting the same issues! Did you find a solution to this?

Negative. I thought of an error with App2 (recently released to production). But then suddenly started happening in App1, which has been stable for several months.

Are you using Supabase or Inngest in your app by any chance?

I really hope someone can shed some light into this matter

Hello !

We’ve been facing the exact same issue for the past few weeks — random HTTP 500s with SIGABRT and SIGSEGV crashes on Vercel (also in eu-central-1). It started suddenly without any code or dependency changes.

After a lot of trial and error, the following changes seem to have stabilized things on our side:

Capture d’écran 2025-10-09 à 15.14.23

Capture d’écran 2025-10-09 à 15.14.06

Capture d’écran 2025-10-09 à 15.14.15

So far, no more crashes since applying these adjustments. I hope this can help you as well!

Hi, thanks for the suggestion. No luck in our case, keep getting the same errors . Will keep trying.

I'm facing the same issue. Are there any updates on this? It’s happening in both FRA1 and CDG1 for me.

I tried the solution above, but it didn’t work.

Is the Vercel team aware of this issue? And to which regions it applies? It’d be great if team Vercel could provide some insight into what’s happening and how to resolve it.

I was getting the same random issues on Vercel (possibly from a mix of Bun and Node environments).

Steps I took:

Removed bun.lock, node_modules, and package-lock.json
Made sure all my package.json scripts use Node commands (removed any bun references)
Matched my local Node version to Vercel’s (20.x)
Locally run npm install and commit updated package.json and package-lock.json
In Vercel → Settings → Build & Development → Framework Settings, set:
- Build Command: npm run build
- Install Command: npm ci
Deploy vercel --force

This setup seems to have resulted in a stable environment for me (for now)

Vercel Alum

Thanks for sharing what worked for you, ! I'll share this with the team internally to make sure that they're aware of it.

Would be keen to hear if this has resolved for you?

Hi . Unfortunately it hasn't been resolved. Appreciate all the suggestions though, but none have worked.

Had the exactly issues. No major changes recently for our code and it started to show up around 9/28. Also encountering something like double free or corruption (out) along with that (SIGABRT / SIGSEGV) error.

We’re seeing the same issue. A mix of Node.js process exited with signal: 11 (SIGSEGV) (core dumped). and Node.js process exited with signal: 6 (SIGABRT) (core dumped).No other logs.

Node 22.20.0
Next 15.2.1
Fluid Compute enabled
Using Prisma, not Supabase

Stumbled upon this suggestion and wondering if it might be related: https://github.com/prisma/prisma/issues/27785#issuecomment-3164426088

GitHub

segfault with debian-bookworm and openssl 3.0.17 · Issue #27785 · prisma/prisma

Important This issue has been fixed in OpenSSL 3.0.17-1~deb12u2 on Debian, please update your OpenSSL version if you are on 3.0.17-1~deb12u1. This issue is still temporarily open for better discove...

Hi . We are using Prisma too, v5.18 for both apps. Would be nice to know if others are using Prisma as well.

I’m on "prisma": "^6.4.1" and not facing the issue anymore.

we’re also on @prisma/client@6.4.1 so its version does not seem to be related.

looking at that thread I linked, this could be affecting many libraries that use OpenSSL, not just Prisma.

Yes, we are on primsa prisma/client: 5.22.0

We’re impacted by this same failure mode.

Fluid compute doesn’t seem to change it.

We’re in iad1.

Prisma 5.22.

Started happening around Oct 15th.

I was using Node 22.x, I switched to 20.x and the problem persists, so I believe the issue isn’t the Node version. I’m using Prisma version 5.14.0, as previously mentioned, it could have something to do with it.

the same

Next.js: 15.2.0
Prisma: 6.4.1
Node: 22

Vercel Staff

Hi and , welcome to the Vercel Community!

I'm sorry you're facing this issue. Is it resolved now?

if not, have you tried the steps here?

Hi . I've tried switching node versions, regenerating the package.lock, and deploying it many times.

Since I first published this problem, App1 (see my original message) it is still getting SIGSEGV errors.

For App2, the only modification I did was changing an await Promise.all to a for loop with await when requesting the DB. I'm aware that using a Promise.all for querying data may not be a best practice if the process triggers many requests at once (even using a connection pool). But in our case, the number of requests it's exactly the same every time (just 5) and the number of users is just one in the preview environment (where we observed the first errors). So I doubt that this change had any implications regarding the root issue

i been getting the same too and cannot understand the root cause Next.js: 15.2.0

Prisma: 6.4.1
Node: 22

actually…. ‘s solution seems to fix issue on our project. I don’t see any issue last 24 hours.

I was skeptical it would resolve the issue but yeah. All I did was removing package-lock.json and recreated it. (it had bunch of changes obviously)

We are still seeing too, and on any route (rsc, api route etc), we’ll try to upgrade all packages to see if it will resolve

Deleting package-lock.json and running npm install again seems to fix this issue.

It didn’t solve the issue for me.

I deployed my app to a competitor, same build command, and no issues.

I solved the issue by upgrading from React 19-RC to React 19, and now everything works perfectly

This topic is closed. It no longer accepts new replies.