Next.js server crash on Linux — recovery guide

TL;DR

A self-hosted Next.js server crashing on Linux leaves your entire application unreachable. Unlike Vercel's managed infrastructure, a crash on your own server means no automatic failover unless you have explicitly configured process management and health monitoring.

Common crash sources

• Uncaught exceptions in getServerSideProps — an unhandled database timeout or API failure in a server-side data fetch brings down the entire Node.js process rather than returning a 500 to that single request
• API route errors — throwing outside a try/catch in a Next.js API route handler crashes the server process
• ISR revalidation failures — when Incremental Static Regeneration fails on a callback and the error propagates uncaught, the revalidation worker can take the process with it
• Middleware infinite loops — a redirect chain in middleware.ts that matches its own destination creates an infinite loop, exhausting the stack

Diagnosing the crash

Check PM2 for crash history and recent logs:

pm2 status
pm2 logs nextapp --lines 200 --err

If the process exited without PM2 catching it, check systemd journal and kernel logs:

journalctl -u pm2-deploy --since "1 hour ago"
dmesg | grep -i "oom\|killed process"

Generate a Node.js crash dump for deeper analysis by setting the environment variable before starting:

NODE_OPTIONS='--abort-on-uncaught-exception' pm2 start ecosystem.config.js

Look for .core or crash report files in the working directory.

Restart strategies

Configure PM2 with exponential backoff to prevent rapid restart loops:

module.exports = {
  apps: [{
    name: 'nextapp',
    script: '.next/standalone/server.js',
    instances: 2,
    exec_mode: 'cluster',
    max_restarts: 10,
    min_uptime: '10s',
    exp_backoff_restart_delay: 100,
    max_memory_restart: '1G',
  }],
};

Graceful shutdown handling

Add a custom server wrapper that handles process signals correctly:

const { createServer } = require('http');
const next = require('next');

const app = next({ dev: false });
const handle = app.getRequestHandler();

app.prepare().then(() => {
  const server = createServer(handle);
  server.listen(3000, () => {
    if (process.send) process.send('ready');
  });

  process.on('SIGINT', () => {
    server.close(() => process.exit(0));
  });
});

Prevention

Wrap all getServerSideProps and API route handlers in try/catch blocks. Use Next.js error boundaries for rendering failures. For ISR, set fallback: 'blocking' and handle revalidation errors explicitly in your data-fetching layer rather than letting them propagate.

Where Reflex helps

Reflex monitors your Next.js process health and detects crash loops before they exhaust PM2's restart limit. When a crash occurs, Reflex captures the error context, executes a recovery playbook — restarting the process, verifying the health endpoint, and rolling back to the previous build if the crash persists — all with a documented audit trail. See How it works.