Solving Cloudflare Worker Resource Limits: A Deep Dive into Performance Optimization

When your production application suddenly starts throwing Error 1102: Worker exceeded resource limits, it's time for some serious performance detective work. This is the story of how we diagnosed, analyzed, and solved a critical resource limit issue in our Next.js blog deployed on Cloudflare Workers.

The Problem: Error 1102 in Production

The Incident

Error 1102 Ray ID: 97444d3e0ece590c • 2025-08-24 16:42:14 UTC
Worker exceeded resource limits

Our blog suddenly became inaccessible, throwing this cryptic error. Users couldn't access any pages, and the Worker was consistently hitting resource limits.

Understanding Cloudflare Worker Limits

Cloudflare Workers have strict resource constraints:

Resource	Free Plan	Paid Plan
CPU Time	10ms	50ms
Memory	128MB	128MB
Bundle Size	1MB	10MB
Subrequests	50	1000

When any of these limits are exceeded, you get Error 1102.

Root Cause Analysis

Initial Investigation

Let's examine what was happening in our application:

1# Check bundle size
2ls -lh src/data/posts.json
3# Output: -rw-r--r-- 1 staff 226K Aug 24 13:33 posts.json

226KB for a single JSON file! This was our first red flag.

Memory Usage Pattern Analysis

Here's what was happening on every request:

The Data Structure Problem

Our posts.json contained:

1{
2  "posts": [
3    {
4      "id": "01-creating-nextjs-project",
5      "slug": "creating-nextjs-project...",
6      "title": "Creating a Next.js Project",
7      "excerpt": "Brief description...",
8      "content": "# Very long content with Mermaid diagrams, code blocks, etc..."
9    }
10    // ... 18 more posts with full content
11  ]
12}

Problem: Every page load imported the entire 226KB file, even when only metadata was needed.

Performance Impact Analysis

Metric	Impact	Consequence
Bundle Size	226KB loaded on import	Slow Worker startup
Memory Usage	All posts in memory	High memory pressure
CPU Time	JSON parsing + processing	Exceeded 10ms limit
Network	Large bundle transfer	Increased latency

The Solution: Lazy Loading Architecture

Strategy Overview

We implemented a two-tier data architecture:

1. Metadata Layer: Lightweight index (10.9KB)
2. Content Layer: Individual files loaded on-demand

Implementation Deep Dive

Step 1: Data Separation Script

1// scripts/optimize-for-workers.js
2async function optimizeForWorkers() {
3  const postsData = JSON.parse(fs.readFileSync(POSTS_FILE, 'utf8'));
4
5  // Create lightweight metadata
6  const metadata = {
7    posts: postsData.posts.map(post => ({
8      id: post.id,
9      slug: post.slug,
10      title: post.title,
11      excerpt: post.excerpt,
12      date: post.date,
13      author: post.author,
14      // Remove content field
15    })),
16    slugs: postsData.slugs,
17    generatedAt: postsData.generatedAt
18  };
19
20  // Save individual content files
21  for (const post of postsData.posts) {
22    const contentFile = path.join(CONTENT_DIR, `${post.slug}.json`);
23    fs.writeFileSync(contentFile, JSON.stringify({ content: post.content }));
24  }
25
26  fs.writeFileSync(METADATA_FILE, JSON.stringify(metadata, null, 2));
27}

Step 2: Optimized Data Loading

1// src/lib/posts-static.ts
2import postsMetadata from '@/data/posts-metadata.json'; // Only 10.9KB!
3
4const contentCache = new Map<string, string>();
5
6async function loadPostContent(slug: string): Promise<string> {
7  if (contentCache.has(slug)) {
8    return contentCache.get(slug)!;
9  }
10
11  try {
12    // Dynamic import - only loads when needed
13    const contentModule = await import(`@/data/content/${slug}.json`);
14    const content = contentModule.content || '';
15
16    contentCache.set(slug, content);
17    return content;
18  } catch (error) {
19    console.error(`Error loading content for ${slug}:`, error);
20    return '';
21  }
22}
23
24// Metadata only - fast and lightweight
25export function getAllPosts(): Omit<Post, 'content'>[] {
26  return postsMetadata.posts.map(post => ({
27    id: post.id,
28    slug: post.slug,
29    title: post.title,
30    date: post.date,
31    author: post.author,
32    excerpt: post.excerpt,
33  }));
34}
35
36// Full post with content - loaded on demand
37export async function getPostBySlug(slug: string): Promise<Post | undefined> {
38  const postMeta = postsMetadata.posts.find(post => post.slug === slug);
39
40  if (!postMeta) return undefined;
41
42  const content = await loadPostContent(slug);
43  return { ...postMeta, content };
44}

New Request Flow

Performance Analysis

Bundle Size Optimization

1# Before optimization
2Original size: 226.3KB
3
4# After optimization
5Metadata size: 10.9KB
6Space saved: 95.2% (215.4KB)

Memory Usage Comparison

Scenario	Before	After	Improvement
Posts listing	226KB	10.9KB	95.2% reduction
Single post	226KB	10.9KB + ~12KB	~90% reduction
Multiple posts	226KB	10.9KB + (n × ~12KB)	Scales linearly

CPU Time Analysis

Network Bandwidth Analysis

Initial Bundle Transfer

Component	Before	After	Savings
Metadata	226KB	10.9KB	215.1KB
Content	Included	On-demand	Variable
Total Initial	226KB	10.9KB	95.2%

Runtime Loading Patterns

1// Posts listing: Only metadata needed
2const posts = getAllPosts(); // 10.9KB loaded
3
4// Individual post: Metadata + specific content
5const post = await getPostBySlug('my-post'); // 10.9KB + ~12KB

Implementation Challenges & Solutions

Challenge 1: Dynamic Imports in Workers

Problem: Cloudflare Workers have limitations with dynamic imports.

Solution: Use static imports with dynamic paths that are known at build time:

1// This works in Workers
2const contentModule = await import(`@/data/content/${slug}.json`);
3
4// This doesn't work in Workers
5const contentModule = await import(dynamicPath);

Challenge 2: Type Safety

Problem: Posts without content need different types.

Solution: Flexible type definitions:

1interface PostCardProps {
2  post: Omit<Post, 'content'> | Post; // Supports both
3}

Challenge 3: Build Pipeline Integration

Problem: Need to run optimization automatically.

Solution: Integrated build pipeline:

1{
2  "scripts": {
3    "prebuild": "node scripts/generate-posts-data.js && node scripts/optimize-for-workers.js"
4  }
5}

Results & Impact

✅ Deployment Success

1# Successful deployment output
2✨ Success! Uploaded 7 files (66 already uploaded) (1.77 sec)
3Total Upload: 13450.13 KiB / gzip: 2695.83 KiB
4Worker Startup Time: 24 ms
5Deployed next-blog triggers (1.27 sec)
6https://next-blog.rkristelijn.workers.dev

Performance Metrics

Metric	Before	After	Improvement
Error Rate	100% (Error 1102)	0%	✅ Resolved
Bundle Size	226KB	10.9KB	95.2% reduction
Memory Usage	High	Low	~95% reduction
Startup Time	Slow	Fast	Significantly improved
Scalability	Limited	High	Linear scaling

Resource Utilization

Lessons Learned

1. Monitor Bundle Sizes Early

1# Add to CI/CD pipeline
2npm run build | grep "First Load JS"

2. Implement Lazy Loading from Start

Don't load everything upfront. Design for on-demand loading:

1// Good: Load what you need
2const metadata = getPostMetadata();
3
4// Bad: Load everything
5const allData = getAllDataIncludingContent();

3. Use Appropriate Data Structures

1// For listings: Metadata only
2interface PostSummary {
3  id: string;
4  title: string;
5  excerpt: string;
6  date: string;
7}
8
9// For details: Full content
10interface PostDetail extends PostSummary {
11  content: string;
12}

4. Cache Strategically

1// Cache expensive operations
2const contentCache = new Map<string, string>();
3
4// But don't cache everything
5// Cache only frequently accessed content

Best Practices for Cloudflare Workers

1. Bundle Size Management

• Keep initial bundles under 50KB
• Use dynamic imports for large content
• Monitor bundle sizes in CI/CD

2. Memory Optimization

• Load data on-demand
• Implement intelligent caching
• Avoid loading entire datasets

3. CPU Time Management

• Minimize JSON parsing
• Use efficient algorithms
• Profile critical paths

4. Monitoring & Alerting

1// Add performance monitoring
2console.time('operation');
3await expensiveOperation();
4console.timeEnd('operation');
5
6// Monitor in production
7if (process.env.NODE_ENV === 'production') {
8  // Log performance metrics
9}

Future Optimizations

1. Content Compression

1# Gzip content files
2gzip content/*.json
3# Potential 60-80% additional savings

2. Edge Caching

1// Cache content at Cloudflare edge
2const cacheKey = `post-content-${slug}`;
3const cached = await caches.default.match(cacheKey);

3. Incremental Loading

1// Load content sections progressively
2const sections = await loadPostSections(slug);

4. Service Worker Caching

1// Client-side caching for repeat visits
2self.addEventListener('fetch', event => {
3  if (event.request.url.includes('/content/')) {
4    event.respondWith(cacheFirst(event.request));
5  }
6});

Conclusion

The Error 1102 "Worker exceeded resource limits" taught us valuable lessons about performance optimization in serverless environments. By implementing lazy loading and reducing our bundle size by 95.2%, we not only solved the immediate problem but also created a more scalable architecture.

Key Takeaways:

1. Monitor resource usage proactively - Don't wait for errors
2. Design for lazy loading - Load only what you need, when you need it
3. Optimize bundle sizes - Every KB matters in serverless environments
4. Implement intelligent caching - Balance memory usage with performance
5. Test at scale - Resource limits become apparent under load

The solution demonstrates that with careful analysis and strategic optimization, even complex applications can run efficiently within Cloudflare Worker constraints while maintaining excellent performance and user experience.

Our blog now handles traffic smoothly, scales efficiently, and stays well within resource limits - proving that sometimes the best optimization is simply not loading what you don't need.

---

Performance Stats:

• ✅ 95.2% bundle size reduction (226KB → 10.9KB)
• ✅ Zero Error 1102 incidents since optimization
• ✅ Linear scalability for additional content
• ✅ Sub-5ms CPU time for most operations

The complete optimization code and scripts are available in our GitHub repository.