Most Next.js applications handle the happy path well. A request comes in, data loads, components render, user sees the page. Error handling is where applications often reveal their real quality — and where App Router introduces some nuances worth understanding.
Here's how error boundaries work in App Router, what the error.js file actually does, and the patterns that make failure handling feel intentional rather than afterthought. This is the approach I use in production for a free AI image generator for beginners where graceful degradation matters more than on sites with simpler data flows.
What Changes in App Router
Pages Router had _error.js and getInitialProps for error handling. App Router introduces a different model built on React's Error Boundary concept, with error.js files that can be nested at any level of the route hierarchy.
The key mental model shift: errors are contained at the nearest error.js boundary, not propagated to a global handler. This means you can have different error UIs for different sections of your app.
The error.js File
Place an error.js file in any route segment to catch errors in that segment and its children:
app/
├── layout.js # Root layout
├── error.js # Catches errors in root segment
├── page.js
├── dashboard/
│ ├── error.js # Catches errors in dashboard only
│ ├── layout.js
│ └── page.js
└── blog/
├── error.js # Catches errors in blog only
└── [slug]/
└── page.js
// app/error.js
'use client'; // Error components must be Client Components
import { useEffect } from 'react';
export default function Error({ error, reset }) {
useEffect(() => {
// Log to error reporting service
console.error(error);
}, [error]);
return (
<div className="flex flex-col items-center justify-center
min-h-[400px] gap-4 p-6 text-center">
<h2 className="text-xl font-semibold text-foreground">
Something went wrong
</h2>
<p className="text-sm text-muted max-w-sm">
{error.message || 'An unexpected error occurred.'}
</p>
<button
onClick={reset}
className="px-4 py-2 bg-orange-500 text-white
rounded-full text-sm font-medium
hover:bg-orange-600 transition-colors"
>
Try again
</button>
</div>
);
}
Important: error.js must be a Client Component ('use client'). React Error Boundaries are a client-side concept — server errors get converted to client-side error events before reaching the boundary.
The reset function retries the failed render. Use it for transient errors (network issues, temporary service failures) rather than permanent ones.
The reset Function — When to Show It
The reset button only makes sense for errors that might succeed on retry:
export default function Error({ error, reset }) {
// Check if error is retryable
const isRetryable = error.digest !== 'NEXT_NOT_FOUND';
return (
<div className="error-container">
<h2>Something went wrong</h2>
<p>{getErrorMessage(error)}</p>
{isRetryable && (
<button onClick={reset}>Try again</button>
)}
{!isRetryable && (
<a href="/">Return home</a>
)}
</div>
);
}
function getErrorMessage(error) {
if (error.message?.includes('fetch')) {
return 'Failed to load data. Check your connection.';
}
if (error.message?.includes('timeout')) {
return 'Request timed out. Please try again.';
}
return 'An unexpected error occurred.';
}
Global Error Handling — global-error.js
The root error.js doesn't catch errors in the root layout.js. For that, you need global-error.js:
// app/global-error.js
'use client';
export default function GlobalError({ error, reset }) {
return (
<html>
<body>
<div className="global-error">
<h1>Application Error</h1>
<p>Something went wrong at the application level.</p>
<button onClick={reset}>Reload</button>
</div>
</body>
</html>
);
}
global-error.js replaces the root layout when it renders, so it needs to include <html> and <body> tags. This is the last-resort catch for errors that escape all other boundaries.
Server Component Error Handling
Server components can throw errors directly, but handling them well requires a pattern slightly different from what you might expect:
// app/blog/[slug]/page.js
async function getBlogPost(slug) {
const response = await fetch(`/api/posts/${slug}`);
if (response.status === 404) {
notFound(); // Triggers not-found.js, not error.js
}
if (!response.ok) {
throw new Error(`Failed to fetch post: ${response.status}`);
// This triggers error.js
}
return response.json();
}
export default async function BlogPost({ params }) {
const post = await getBlogPost(params.slug);
return <PostContent post={post} />;
}
notFound() vs throw Error(): Use notFound() for expected missing resources — it renders not-found.js and returns a 404 status. Use throw for unexpected failures — it renders error.js.
The not-found.js File
For 404-style errors, not-found.js provides a cleaner separation than error.js:
// app/not-found.js
import Link from 'next/link';
export default function NotFound() {
return (
<div className="flex flex-col items-center justify-center
min-h-[400px] gap-4 text-center p-6">
<h2 className="text-2xl font-semibold">Page not found</h2>
<p className="text-muted text-sm">
The page you're looking for doesn't exist or has been moved.
</p>
<Link
href="/"
className="px-4 py-2 bg-foreground text-background
rounded-full text-sm font-medium"
>
Return home
</Link>
</div>
);
}
Error Boundaries for Async Operations in Client Components
Client components that fetch data need their own error handling since they're outside the server component error flow:
'use client';
import { useState, useEffect } from 'react';
export function DataFetcher({ url }) {
const [data, setData] = useState(null);
const [error, setError] = useState(null);
const [loading, setLoading] = useState(true);
useEffect(() => {
async function fetchData() {
try {
setLoading(true);
setError(null);
const response = await fetch(url);
if (!response.ok) throw new Error(`HTTP ${response.status}`);
const json = await response.json();
setData(json);
} catch (err) {
setError(err.message);
} finally {
setLoading(false);
}
}
fetchData();
}, [url]);
if (loading) return <Skeleton />;
if (error) return (
<div className="error-inline">
<p className="text-sm text-red-500">{error}</p>
<button
onClick={() => fetchData()}
className="text-xs text-muted hover:text-foreground"
>
Retry
</button>
</div>
);
return <DataDisplay data={data} />;
}
Error Reporting Integration
error.js is the right place to integrate with error monitoring:
'use client';
import { useEffect } from 'react';
export default function Error({ error, reset }) {
useEffect(() => {
// Send to your error monitoring service
reportError({
message: error.message,
digest: error.digest, // Next.js server error ID
stack: error.stack,
timestamp: new Date().toISOString(),
path: window.location.pathname,
});
}, [error]);
return (
// Error UI
);
}
async function reportError(errorData) {
try {
await fetch('/api/errors', {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify(errorData),
});
} catch {
// Don't let error reporting itself cause more errors
console.error('Failed to report error:', errorData);
}
}
The error.digest is particularly useful — it's a hash that Next.js uses to identify server-side errors in logs without exposing sensitive stack traces to the client.
The Error Boundary Hierarchy in Practice
For a real application, the error boundary structure might look like:
global-error.js ← Last resort, includes <html>
├── error.js ← Root-level errors
│ ├── (auth)/
│ │ └── error.js ← Auth-specific errors
│ ├── dashboard/
│ │ └── error.js ← Dashboard-specific errors
│ └── api/
│ └── route.js ← API routes handle their own errors
This lets you show relevant error messages — "Your session expired, please log in again" for auth errors, "Failed to load dashboard data" for dashboard errors — rather than a generic "something went wrong" for everything.
What I Built This On
This error handling pattern runs in production on pixova.io. The generation pipeline has several async steps where things can go wrong — API timeouts, inference failures, upload errors — and the error boundary setup means failures at each step show appropriate messages rather than breaking the whole page.
The reset button is particularly important for generation tools where users are mid-workflow when something fails — it retries without losing the prompt they just wrote.
Questions on specific error scenarios? Drop them in the comments.
Top comments (1)
This post is essentially about building a “Bharatiya Ephemeris” derived from classical Indian astronomical/astrological texts and benchmarking it against NASA JPL (DE/DE441-type) ephemerides.
What the idea is really about
At its core, this kind of project sits in a very specific intersection:
NASA JPL ephemerides (DE440/DE441) → modern gold-standard numerical integration of solar system motion
Classical texts (Sūrya Siddhānta, etc.) → pre-modern mathematical models of planetary motion
Benchmarking layer → comparing “text-based computed positions” vs high-precision physical simulation
Modern ephemerides like JPL are used for spacecraft navigation and are accurate to sub-arcsecond precision, built from observational data and numerical integration of gravitational models . That’s fundamentally different from classical astronomical models, which rely on geometric/periodic approximations rather than full-body gravitational simulation.
Why this is interesting (and hard)
The interesting part of this project is not just cultural — it’s technical:
Classical models are algorithmic, not observational
They approximate planetary motion using periodic functions and corrections.
They were designed for naked-eye astronomy, not milliarcsecond precision.
JPL is physics-first, not theory-first
Built from decades of observational data + numerical integration.
Continuously refined for mission-critical navigation.
So the benchmark is asymmetrical by design
You’re comparing:
symbolic/approximate mathematical systems
vs
data-fitted physical simulation of the solar system
That means any deviation you find is expected — but the size and pattern of deviation is where the real insight lives.
The real value of projects like this
Even if the classical model doesn’t “match” JPL numerically, this kind of work can still produce meaningful outcomes:
understanding how ancient algorithms approximate periodic motion
identifying where classical corrections diverge most (planet, epoch, latitude, etc.)
building hybrid models (classical structure + modern ephemeris backend)
or creating reproducible historical simulation tools
The key takeaway
This isn’t really a “who is more accurate” problem.
It’s more like:
classical astronomy is a model of meaning and approximation, while JPL ephemerides are a model of physical reality
So benchmarking them is less about correctness and more about mapping two fundamentally different world models onto the same coordinate system.
Projects like this are valuable mainly when they stay honest about that distinction 🤝