akornmeier/cloudflare

Cloudflare Developer Platform Skill

Cloudflare Developer Platform is a comprehensive edge computing ecosystem for building full-stack applications on Cloudflare's global network. It includes serverless functions, databases, storage, AI/ML capabilities, and static site hosting.

When to Use This Skill

Use this skill when:

• Building serverless applications on the edge
• Implementing edge databases (D1 SQLite)
• Working with object storage (R2) or key-value stores (KV)
• Creating real-time applications with WebSockets (Durable Objects)
• Integrating AI/ML capabilities (Workers AI, AI Gateway, Agents)
• Deploying static sites with serverless functions (Pages)
• Building full-stack applications with frameworks (Next.js, Remix, Astro, etc.)
• Implementing message queues and background jobs (Queues)
• Optimizing for global performance and low latency

Core Concepts

Edge Computing Platform

Cloudflare's Edge Network: Code runs on servers globally distributed across 300+ cities, executing requests from the nearest location for ultra-low latency.

Key Components:

• Workers: Serverless functions on the edge
• D1: SQLite database with global read replication
• KV: Distributed key-value store with eventual consistency
• R2: Object storage with zero egress fees
• Durable Objects: Stateful compute with WebSocket support
• Queues: Message queue system for async processing
• Pages: Static site hosting with serverless functions
• Workers AI: Run AI models on the edge
• AI Gateway: Unified interface for AI providers

Execution Model

V8 Isolates: Lightweight execution environments (faster than containers) with:

• Millisecond cold starts
• Zero infrastructure management
• Automatic scaling
• Pay-per-request pricing

Handler Types:

• fetch: HTTP requests
• scheduled: Cron jobs
• queue: Message processing
• tail: Log aggregation
• email: Email handling
• alarm: Durable Object timers

Getting Started with Workers

Installation

# Install Wrangler CLI
npm install -g wrangler

# Login to Cloudflare
wrangler login

# Create new project
wrangler init my-worker
cd my-worker

# Start local development
wrangler dev

# Deploy to production
wrangler deploy

Basic Worker

// src/index.ts
export default {
  async fetch(request: Request, env: Env): Promise<Response> {
    return new Response('Hello from Cloudflare Workers!');
  },
};

Configuration (wrangler.toml)

name = "my-worker"
main = "src/index.ts"
compatibility_date = "2024-01-01"

# Environment variables
[vars]
ENVIRONMENT = "production"

# Bindings (added per product below)

Language Support

• JavaScript/TypeScript: Primary language (full Node.js compatibility)
• Python: Beta support via Workers Python
• Rust: Compile to WebAssembly

Storage Products

D1 (SQLite Database)

Use Cases: Relational data, complex queries, ACID transactions

Setup:

# Create database
wrangler d1 create my-database

# Add to wrangler.toml
[[d1_databases]]
binding = "DB"
database_name = "my-database"
database_id = "YOUR_DATABASE_ID"

# Generate and apply schema
wrangler d1 execute my-database --file=./schema.sql

Usage:

export default {
  async fetch(request: Request, env: Env) {
    // Query
    const result = await env.DB.prepare('SELECT * FROM users WHERE id = ?').bind(userId).first();

    // Insert
    await env.DB.prepare('INSERT INTO users (name, email) VALUES (?, ?)')
      .bind('Alice', '[email protected]')
      .run();

    // Batch (atomic)
    await env.DB.batch([
      env.DB.prepare('UPDATE accounts SET balance = balance - 100 WHERE id = ?').bind(user1),
      env.DB.prepare('UPDATE accounts SET balance = balance + 100 WHERE id = ?').bind(user2),
    ]);

    return new Response(JSON.stringify(result));
  },
};

Key Features:

• Global read replication (low-latency reads)
• Single-writer consistency
• Standard SQLite syntax
• 25GB database size limit

KV (Key-Value Store)

Use Cases: Cache, sessions, feature flags, rate limiting

Setup:

# Create namespace
wrangler kv:namespace create MY_KV

# Add to wrangler.toml
[[kv_namespaces]]
binding = "KV"
id = "YOUR_NAMESPACE_ID"

Usage:

export default {
  async fetch(request: Request, env: Env) {
    // Put with TTL
    await env.KV.put('session:token', JSON.stringify(data), {
      expirationTtl: 3600, // 1 hour
    });

    // Get
    const data = await env.KV.get('session:token', 'json');

    // Delete
    await env.KV.delete('session:token');

    // List with prefix
    const list = await env.KV.list({ prefix: 'user:123:' });

    return new Response(JSON.stringify(data));
  },
};

Key Features:

• Sub-millisecond reads (edge-cached)
• Eventual consistency (~60 seconds globally)
• 25MB value size limit
• Automatic expiration (TTL)

R2 (Object Storage)

Use Cases: File storage, media hosting, backups, static assets

Setup:

# Create bucket
wrangler r2 bucket create my-bucket

# Add to wrangler.toml
[[r2_buckets]]
binding = "R2_BUCKET"
bucket_name = "my-bucket"

Usage:

export default {
  async fetch(request: Request, env: Env) {
    // Put object
    await env.R2_BUCKET.put('path/to/file.jpg', fileBuffer, {
      httpMetadata: {
        contentType: 'image/jpeg',
      },
    });

    // Get object
    const object = await env.R2_BUCKET.get('path/to/file.jpg');
    if (!object) {
      return new Response('Not found', { status: 404 });
    }

    // Stream response
    return new Response(object.body, {
      headers: {
        'Content-Type': object.httpMetadata?.contentType || 'application/octet-stream',
      },
    });

    // Delete
    await env.R2_BUCKET.delete('path/to/file.jpg');

    // List
    const list = await env.R2_BUCKET.list({ prefix: 'uploads/' });
  },
};

Key Features:

• S3-compatible API
• Zero egress fees (huge cost advantage)
• Unlimited storage
• 5TB object size limit
• Multipart upload support

Durable Objects

Use Cases: Real-time apps, WebSockets, coordination, stateful logic

Setup:

# wrangler.toml
[[durable_objects.bindings]]
name = "COUNTER"
class_name = "Counter"
script_name = "my-worker"

Usage:

// Define Durable Object class
export class Counter {
  state: DurableObjectState;

  constructor(state: DurableObjectState, env: Env) {
    this.state = state;
  }

  async fetch(request: Request) {
    // Get current count
    let count = (await this.state.storage.get<number>('count')) || 0;

    // Increment
    count++;
    await this.state.storage.put('count', count);

    return new Response(JSON.stringify({ count }));
  }
}

// Use in Worker
export default {
  async fetch(request: Request, env: Env) {
    // Get Durable Object instance
    const id = env.COUNTER.idFromName('global-counter');
    const counter = env.COUNTER.get(id);

    // Forward request
    return counter.fetch(request);
  },
};

WebSocket Example:

export class ChatRoom {
  state: DurableObjectState;
  sessions: Set<WebSocket>;

  constructor(state: DurableObjectState) {
    this.state = state;
    this.sessions = new Set();
  }

  async fetch(request: Request) {
    const pair = new WebSocketPair();
    const [client, server] = Object.values(pair);

    this.state.acceptWebSocket(server);
    this.sessions.add(server);

    return new Response(null, { status: 101, webSocket: client });
  }

  async webSocketMessage(ws: WebSocket, message: string) {
    // Broadcast to all connected clients
    for (const session of this.sessions) {
      session.send(message);
    }
  }

  async webSocketClose(ws: WebSocket) {
    this.sessions.delete(ws);
  }
}

Key Features:

• Single-instance coordination (strong consistency)
• Persistent storage (1GB limit on paid plans)
• WebSocket support
• Automatic hibernation for inactive objects

Queues

Use Cases: Background jobs, email sending, async processing

Setup:

# wrangler.toml
[[queues.producers]]
binding = "MY_QUEUE"
queue = "my-queue"

[[queues.consumers]]
queue = "my-queue"
max_batch_size = 10
max_batch_timeout = 30

Usage:

// Producer: Send messages
export default {
  async fetch(request: Request, env: Env) {
    await env.MY_QUEUE.send({
      type: 'email',
      to: '[email protected]',
      subject: 'Welcome!',
    });

    return new Response('Message queued');
  },
};

// Consumer: Process messages
export default {
  async queue(batch: MessageBatch<any>, env: Env) {
    for (const message of batch.messages) {
      try {
        await processMessage(message.body);
        message.ack(); // Acknowledge success
      } catch (error) {
        message.retry(); // Retry on failure
      }
    }
  },
};

Key Features:

• At-least-once delivery
• Automatic retries (exponential backoff)
• Dead-letter queue support
• Batch processing

AI Products

Workers AI

Use Cases: Run AI models directly on the edge

Setup:

# wrangler.toml
[ai]
binding = "AI"

Usage:

export default {
  async fetch(request: Request, env: Env) {
    // Text generation
    const response = await env.AI.run('@cf/meta/llama-3-8b-instruct', {
      messages: [{ role: 'user', content: 'What is edge computing?' }],
    });

    // Image classification
    const imageResponse = await env.AI.run('@cf/microsoft/resnet-50', {
      image: imageBuffer,
    });

    // Text embeddings
    const embeddings = await env.AI.run('@cf/baai/bge-base-en-v1.5', {
      text: 'Hello world',
    });

    return new Response(JSON.stringify(response));
  },
};

Available Models:

• LLMs: Llama 3, Mistral, Gemma, Qwen
• Image: Stable Diffusion, DALL-E, ResNet
• Embeddings: BGE, GTE
• Translation, summarization, sentiment analysis

AI Gateway

Use Cases: Unified interface for AI providers with caching, rate limiting, analytics

Setup:

// OpenAI via AI Gateway
const response = await fetch(
  'https://gateway.ai.cloudflare.com/v1/{account_id}/{gateway_id}/openai/chat/completions',
  {
    method: 'POST',
    headers: {
      Authorization: `Bearer ${env.OPENAI_API_KEY}`,
      'Content-Type': 'application/json',
    },
    body: JSON.stringify({
      model: 'gpt-4',
      messages: [{ role: 'user', content: 'Hello!' }],
    }),
  }
);

Features:

• Request caching (reduce costs)
• Rate limiting
• Analytics and logging
• Supports OpenAI, Anthropic, HuggingFace, etc.

Agents

Use Cases: Build AI agents with tools and workflows

import { Agent } from '@cloudflare/agents';

export default {
  async fetch(request: Request, env: Env) {
    const agent = new Agent({
      model: '@cf/meta/llama-3-8b-instruct',
      tools: [
        {
          name: 'get_weather',
          description: 'Get current weather',
          parameters: {
            type: 'object',
            properties: {
              location: { type: 'string' },
            },
          },
          handler: async ({ location }) => {
            // Fetch weather data
            return { temperature: 72, conditions: 'sunny' };
          },
        },
      ],
    });

    const result = await agent.run('What is the weather in San Francisco?');
    return new Response(JSON.stringify(result));
  },
};

AI Search (RAG)

Use Cases: Build retrieval-augmented generation applications

import { VectorizeIndex } from '@cloudflare/workers-types';

export default {
  async fetch(request: Request, env: Env) {
    // Generate embeddings
    const embeddings = await env.AI.run('@cf/baai/bge-base-en-v1.5', {
      text: query,
    });

    // Search vector database
    const results = await env.VECTORIZE_INDEX.query(embeddings.data[0], {
      topK: 5,
    });

    // Generate response with context
    const response = await env.AI.run('@cf/meta/llama-3-8b-instruct', {
      messages: [
        {
          role: 'system',
          content: `Context: ${results.matches.map((m) => m.metadata.text).join('\n')}`,
        },
        { role: 'user', content: query },
      ],
    });

    return new Response(JSON.stringify(response));
  },
};

Cloudflare Pages

Static Sites + Serverless Functions

Deployment:

# Deploy via Git (recommended)
# Connect GitHub repo in Cloudflare dashboard

# Or deploy via CLI
wrangler pages deploy ./dist

Pages Functions

Directory-based routing in functions/:

functions/
├── api/
│   ├── users/
│   │   └── [id].ts       # /api/users/:id
│   └── posts.ts          # /api/posts
└── _middleware.ts        # Global middleware

Example Function:

// functions/api/users/[id].ts
export async function onRequestGet(context) {
  const { params, env } = context;
  const user = await env.DB.prepare('SELECT * FROM users WHERE id = ?').bind(params.id).first();

  return new Response(JSON.stringify(user), {
    headers: { 'Content-Type': 'application/json' },
  });
}

Middleware:

// functions/_middleware.ts
export async function onRequest(context) {
  const start = Date.now();
  const response = await context.next();
  const duration = Date.now() - start;

  console.log(`${context.request.method} ${context.request.url} - ${duration}ms`);
  return response;
}

Framework Support

Next.js:

npx create-next-app@latest my-app
cd my-app
npm install -D @cloudflare/next-on-pages
npx @cloudflare/next-on-pages
wrangler pages deploy .vercel/output/static

Remix:

npx create-remix@latest --template cloudflare/remix

Astro:

npm create astro@latest
# Select "Cloudflare" adapter during setup

SvelteKit:

npm create svelte@latest
npm install -D @sveltejs/adapter-cloudflare

Wrangler CLI Essentials

Core Commands

# Development
wrangler dev                    # Local development server
wrangler dev --remote          # Dev on real Cloudflare infrastructure

# Deployment
wrangler deploy                # Deploy to production
wrangler deploy --dry-run     # Preview changes without deploying

# Logs
wrangler tail                  # Real-time logs
wrangler tail --format pretty # Formatted logs

# Versions
wrangler deployments list      # List deployments
wrangler rollback [version]   # Rollback to previous version

# Secrets
wrangler secret put SECRET_NAME    # Add secret
wrangler secret list               # List secrets
wrangler secret delete SECRET_NAME # Delete secret

Project Management

# Create projects
wrangler init my-worker        # Create Worker
wrangler pages project create  # Create Pages project

# Database
wrangler d1 create my-db           # Create D1 database
wrangler d1 execute my-db --file=schema.sql
wrangler d1 execute my-db --command="SELECT * FROM users"

# KV
wrangler kv:namespace create MY_KV
wrangler kv:key put --binding=MY_KV "key" "value"
wrangler kv:key get --binding=MY_KV "key"

# R2
wrangler r2 bucket create my-bucket
wrangler r2 object put my-bucket/file.txt --file=./file.txt

Integration Patterns

Full-Stack Application Architecture

┌─────────────────────────────────────────┐
│         Cloudflare Pages (Frontend)      │
│    Next.js / Remix / Astro / SvelteKit  │
└──────────────────┬──────────────────────┘
                   │
┌──────────────────▼──────────────────────┐
│      Workers (API Layer / BFF)          │
│    - Routing                             │
│    - Authentication                      │
│    - Business logic                      │
└─┬──────┬──────┬──────┬──────┬───────────┘
  │      │      │      │      │
  ▼      ▼      ▼      ▼      ▼
┌────┐ ┌────┐ ┌────┐ ┌────┐ ┌────────────┐
│ D1 │ │ KV │ │ R2 │ │ DO │ │ Workers AI │
└────┘ └────┘ └────┘ └────┘ └────────────┘

Polyglot Storage Pattern

export default {
  async fetch(request: Request, env: Env) {
    const url = new URL(request.url);

    // KV: Fast cache
    const cached = await env.KV.get(url.pathname);
    if (cached) return new Response(cached);

    // D1: Structured data
    const user = await env.DB.prepare('SELECT * FROM users WHERE id = ?').bind(userId).first();

    // R2: Media files
    const avatar = await env.R2_BUCKET.get(`avatars/${user.id}.jpg`);

    // Durable Objects: Real-time coordination
    const chat = env.CHAT_ROOM.get(env.CHAT_ROOM.idFromName(roomId));

    // Queue: Async processing
    await env.EMAIL_QUEUE.send({ to: user.email, template: 'welcome' });

    return new Response(JSON.stringify({ user, avatar }));
  },
};

Authentication Pattern

import { verifyJWT, createJWT } from './jwt';

export default {
  async fetch(request: Request, env: Env) {
    const url = new URL(request.url);

    // Login
    if (url.pathname === '/api/login') {
      const { email, password } = await request.json();

      const user = await env.DB.prepare('SELECT * FROM users WHERE email = ?').bind(email).first();

      if (!user || !(await verifyPassword(password, user.password_hash))) {
        return new Response('Invalid credentials', { status: 401 });
      }

      const token = await createJWT({ userId: user.id }, env.JWT_SECRET);

      return new Response(JSON.stringify({ token }), {
        headers: { 'Content-Type': 'application/json' },
      });
    }

    // Protected route
    const authHeader = request.headers.get('Authorization');
    if (!authHeader) {
      return new Response('Unauthorized', { status: 401 });
    }

    const token = authHeader.replace('Bearer ', '');
    const payload = await verifyJWT(token, env.JWT_SECRET);

    // Store session in KV
    await env.KV.put(`session:${payload.userId}`, JSON.stringify(payload), {
      expirationTtl: 86400, // 24 hours
    });

    return new Response('Authenticated');
  },
};

Cache Strategy

export default {
  async fetch(request: Request, env: Env) {
    const cache = caches.default;
    const cacheKey = new Request(request.url);

    // Check cache
    let response = await cache.match(cacheKey);
    if (response) return response;

    // Check KV (distributed cache)
    const kvCached = await env.KV.get(request.url);
    if (kvCached) {
      response = new Response(kvCached);
      await cache.put(cacheKey, response.clone());
      return response;
    }

    // Fetch from origin (D1, R2, etc.)
    const data = await fetchFromOrigin(request, env);
    response = new Response(data);

    // Store in both caches
    await cache.put(cacheKey, response.clone());
    await env.KV.put(request.url, data, { expirationTtl: 3600 });

    return response;
  },
};

Best Practices

Performance

1. Minimize Cold Starts: Keep Workers lightweight (<1MB bundled)
2. Use Bindings Over Fetch: Direct bindings are faster than HTTP calls
3. Edge Caching: Leverage KV and Cache API for frequently accessed data
4. Batch Operations: Use D1 batch for multiple queries
5. Stream Large Responses: Use Response.body streams for large files

Security

1. Secrets Management: Use wrangler secret for API keys
2. Environment Isolation: Separate production/staging/development
3. Input Validation: Sanitize user input
4. Rate Limiting: Use KV or Durable Objects for rate limiting
5. CORS: Configure proper CORS headers

Cost Optimization

1. R2 for Large Files: Zero egress fees vs S3
2. KV for Caching: Reduce D1/R2 requests
3. Request Deduplication: Cache identical requests
4. Efficient Queries: Index D1 tables properly
5. Monitor Usage: Use Cloudflare Analytics

Development Workflow

1. Local Development: Use wrangler dev for testing
2. Type Safety: Use TypeScript with @cloudflare/workers-types
3. Testing: Use Vitest with unstable_dev()
4. CI/CD: GitHub Actions with cloudflare/wrangler-action
5. Gradual Deployments: Use percentage-based rollouts

Common Patterns

API Gateway

import { Hono } from 'hono';

const app = new Hono();

app.get('/api/users/:id', async (c) => {
  const user = await c.env.DB.prepare('SELECT * FROM users WHERE id = ?')
    .bind(c.req.param('id'))
    .first();

  return c.json(user);
});

app.post('/api/users', async (c) => {
  const { name, email } = await c.req.json();

  await c.env.DB.prepare('INSERT INTO users (name, email) VALUES (?, ?)').bind(name, email).run();

  return c.json({ success: true }, 201);
});

export default app;

Image Transformation

export default {
  async fetch(request: Request, env: Env) {
    const url = new URL(request.url);
    const imageKey = url.pathname.replace('/images/', '');

    // Get from R2
    const object = await env.R2_BUCKET.get(imageKey);
    if (!object) {
      return new Response('Not found', { status: 404 });
    }

    // Transform with Cloudflare Images
    return new Response(object.body, {
      headers: {
        'Content-Type': object.httpMetadata?.contentType || 'image/jpeg',
        'Cache-Control': 'public, max-age=86400',
        'cf-image-resize': JSON.stringify({
          width: 800,
          height: 600,
          fit: 'cover',
        }),
      },
    });
  },
};

Rate Limiting (KV)

async function rateLimit(ip: string, env: Env): Promise<boolean> {
  const key = `ratelimit:${ip}`;
  const limit = 100; // requests per minute
  const window = 60; // seconds

  const current = await env.KV.get(key);
  const count = current ? parseInt(current) : 0;

  if (count >= limit) {
    return false; // Rate limit exceeded
  }

  await env.KV.put(key, (count + 1).toString(), {
    expirationTtl: window,
  });

  return true;
}

export default {
  async fetch(request: Request, env: Env) {
    const ip = request.headers.get('CF-Connecting-IP') || 'unknown';

    if (!(await rateLimit(ip, env))) {
      return new Response('Rate limit exceeded', { status: 429 });
    }

    return new Response('OK');
  },
};

Scheduled Jobs

# wrangler.toml
[triggers]
crons = ["0 0 * * *"] # Daily at midnight

export default {
  async scheduled(event: ScheduledEvent, env: Env) {
    // Cleanup old sessions
    const sessions = await env.KV.list({ prefix: 'session:' });
    for (const key of sessions.keys) {
      const session = await env.KV.get(key.name, 'json');
      if (session.expiresAt < Date.now()) {
        await env.KV.delete(key.name);
      }
    }
  },
};

Troubleshooting

Common Issues

"Module not found" errors

• Ensure dependencies are in package.json
• Run npm install before deploying
• Check compatibility_date in wrangler.toml

Database connection errors (D1)

• Verify database_id in wrangler.toml
• Check database exists: wrangler d1 list
• Run migrations: wrangler d1 execute DB --file=schema.sql

KV not found errors

• Create namespace: wrangler kv:namespace create MY_KV
• Add binding to wrangler.toml
• Deploy after configuration changes

Cold start timeout

• Reduce bundle size (<1MB ideal)
• Remove unnecessary dependencies
• Use dynamic imports for large libraries

CORS errors

• Add CORS headers to responses:

  return new Response(data, {
    headers: {
      'Access-Control-Allow-Origin': '*',
      'Access-Control-Allow-Methods': 'GET, POST, OPTIONS',
      'Access-Control-Allow-Headers': 'Content-Type',
    },
  });
  

Deployment fails

• Check wrangler version: wrangler --version
• Verify authentication: wrangler whoami
• Review build errors in console output

Debugging

# Real-time logs
wrangler tail

# Local debugging with breakpoints
wrangler dev --local

# Remote debugging
wrangler dev --remote

# Check deployment status
wrangler deployments list

Decision Matrix

| Need | Choose | | --------------------- | --------------- | | Sub-millisecond reads | KV | | SQL queries | D1 | | Large files (>25MB) | R2 | | Real-time WebSockets | Durable Objects | | Async background jobs | Queues | | ACID transactions | D1 | | Strong consistency | Durable Objects | | Zero egress costs | R2 | | AI inference | Workers AI | | Static site hosting | Pages | | Serverless functions | Workers | | Multi-provider AI | AI Gateway |

Framework-Specific Guides

Next.js

• Use @cloudflare/next-on-pages adapter
• Configure next.config.js for edge runtime
• Deploy via wrangler pages deploy

Remix

• Use official Cloudflare template
• Configure server.ts for Workers
• Access bindings via context.cloudflare.env

Astro

• Use @astrojs/cloudflare adapter
• Enable SSR in astro.config.mjs
• Access env via Astro.locals.runtime.env

SvelteKit

• Use @sveltejs/adapter-cloudflare
• Configure in svelte.config.js
• Access platform via event.platform.env

Resources

• Documentation: https://developers.cloudflare.com
• Wrangler CLI: https://developers.cloudflare.com/workers/wrangler/
• Discord Community: https://discord.cloudflare.com
• Examples: https://developers.cloudflare.com/workers/examples/
• GitHub: https://github.com/cloudflare
• Status Page: https://www.cloudflarestatus.com

Implementation Checklist

Workers Setup

• [ ] Install Wrangler CLI (npm install -g wrangler)
• [ ] Login to Cloudflare (wrangler login)
• [ ] Create project (wrangler init)
• [ ] Configure wrangler.toml
• [ ] Add environment variables/secrets
• [ ] Test locally (wrangler dev)
• [ ] Deploy (wrangler deploy)

Storage Setup (as needed)

• [ ] Create D1 database and apply schema
• [ ] Create KV namespace
• [ ] Create R2 bucket
• [ ] Configure Durable Objects
• [ ] Set up Queues
• [ ] Add bindings to wrangler.toml

Pages Setup

• [ ] Connect Git repository or use CLI
• [ ] Configure build settings
• [ ] Set environment variables
• [ ] Add Pages Functions (if needed)
• [ ] Deploy and test

Production Checklist

• [ ] Set up custom domain
• [ ] Configure DNS records
• [ ] Enable SSL/TLS
• [ ] Set up monitoring/analytics
• [ ] Configure rate limiting
• [ ] Implement error handling
• [ ] Set up CI/CD pipeline
• [ ] Test gradual deployments
• [ ] Document rollback procedure
• [ ] Configure logging/observability