Your CI is slow. Your pre-commit hooks timeout. Developers disable linting to ship faster.
The culprit? eslint-plugin-import.
The Performance Gap
┌─────────────────────────────────────────────────────┐
│ Linting 10,000 files │
├─────────────────────────────────────────────────────┤
│ eslint-plugin-import: 45.0s ███████████████████│
│ eslint-plugin-import-next: 0.4s ▏ │
└─────────────────────────────────────────────────────┘
That's 100x faster. Not a typo.
Why Is It So Slow?
1. Cold Module Resolution
// eslint-plugin-import resolves EVERY import from scratch
import { Button } from '@company/ui'; // Resolves entire package
// On every lint run. Every file. Every import.
2. The
no-cycle Problem
Click to see why no-cycle is a performance killer
The import/no-cycle rule builds a complete dependency graph.
For N files with M imports each:
- Time complexity: O(N × M²)
- Memory: Entire graph in RAM
- Result: OOM on large monorepos
# Real GitHub issues:
# "import/no-cycle takes 70% of lint time" (#2182)
# "OOM checking circular dependencies"
# "Minutes to lint a monorepo"
3. No Caching
Every lint run repeats the same work. No incremental analysis.
The Solution
We rebuilt module resolution with:
| Feature | eslint-plugin-import | eslint-plugin-import-next |
|---|---|---|
| Caching | ❌ None | ✅ Cross-file shared cache |
| Cycle Detection | O(N × M²) | O(N) with memoization |
| TypeScript | 🐌 Slow resolver | ⚡ Native TS support |
| Flat Config | ⚠️ Partial | ✅ Native |
Quick Migration
npm uninstall eslint-plugin-import
npm install --save-dev eslint-plugin-import-next
// eslint.config.js
import importNext from 'eslint-plugin-import-next';
export default [importNext.configs.recommended];
That's it. Same rules, 100x faster.
Benchmark It Yourself
⭐ Star on GitHub and Run the Benchmark
# Compare on your own codebase
time npx eslint --no-cache . # With eslint-plugin-import
time npx eslint --no-cache . # With eslint-plugin-import-next
📦 npm: eslint-plugin-import-next
📖 Migration Guide
🚀 Is your CI slow? Drop a comment with your lint times!
Follow for more performance deep-dives:
GitHub | LinkedIn