cloudflare-workers

开发编程需配置 API Key

使用 Cloudflare Workers 快速开发——在 Cloudflare 全球网络上构建并部署无服务器应用。适用于构建 API、全栈 Web 应用...

后端开发

517

下载量

安装量

收藏量

★ 5.0

综合评分ⓘ

概述

更多信息

Cloudflare Workers#

Overview#

Cloudflare Workers is a serverless execution environment that runs JavaScript, TypeScript, Python, and Rust code on Cloudflare's global network. Workers execute in milliseconds, scale automatically, and integrate with Cloudflare's storage and compute products through bindings.

Key Benefits:

Zero cold starts - Workers run in V8 isolates, not containers
Global deployment - Code runs in 300+ cities worldwide
Rich ecosystem - Bindings to D1, KV, R2, Durable Objects, Queues, Containers, Workflows, and more
Full-stack capable - Build APIs and serve static assets in one project
Standards-based - Uses Web APIs (fetch, crypto, streams, WebSockets)

When to Use This Skill#

Use Cloudflare Workers for:

APIs and backends - RESTful APIs, GraphQL, tRPC, WebSocket servers
Full-stack applications - React, Next.js, Remix, Astro, Vue, Svelte with static assets
Edge middleware - Authentication, rate limiting, A/B testing, routing
Background processing - Scheduled jobs (cron), queue consumers, webhooks
Data transformation - ETL pipelines, real-time data processing
AI applications - RAG systems, chatbots, image generation with Workers AI
Durable workflows - Multi-step long-running tasks with automatic retries (Workflows)
Container workloads - Run Docker containers alongside Workers (Containers)
MCP servers - Host remote Model Context Protocol servers
Proxy and gateway - API gateways, content transformation, protocol translation

Quick Start Workflow#

1. Install Wrangler CLI#

bash

npm install -g wrangler

# Login to Cloudflare
wrangler login

2. Create a New Worker#

bash

# Using C3 (create-cloudflare) - recommended
npm create cloudflare@latest my-worker

# Or create manually
wrangler init my-worker
cd my-worker

3. Write Your Worker#

Basic HTTP API (TypeScript):

typescript

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

    if (url.pathname === "/api/hello") {
      return Response.json({ message: "Hello from Workers!" });
    }

    return new Response("Not found", { status: 404 });
  },
};

With environment variables and KV:

typescript

interface Env {
  MY_VAR: string;
  MY_KV: KVNamespace;
}

export default {
  async fetch(request: Request, env: Env): Promise<Response> {
    // Access environment variable
    const greeting = env.MY_VAR;

    // Read from KV
    const value = await env.MY_KV.get("my-key");

    return Response.json({ greeting, value });
  },
};

4. Develop Locally#

bash

# Start local development server with hot reload
wrangler dev

# Access at http://localhost:8787

5. Deploy to Production#

bash

# Deploy to workers.dev subdomain
wrangler deploy

# Deploy to custom domain (configure in wrangler.toml)
wrangler deploy

Core Concepts#

Workers Runtime#

Workers use the V8 JavaScript engine with Web Standard APIs:

Execution model: Isolates (not containers) - instant cold starts
CPU time limit: 10ms (Free), 30s (Paid) per request
Memory limit: 128 MB per isolate
Languages: JavaScript, TypeScript, Python, Rust
APIs: fetch, crypto, streams, WebSockets, WebAssembly

Supported APIs:

Fetch API (HTTP requests)
URL API (URL parsing)
Web Crypto (encryption, hashing)
Streams API (data streaming)
WebSockets (real-time communication)
Cache API (edge caching)
HTML Rewriter (HTML transformation)

Handlers#

Workers respond to events through handlers:

Fetch Handler (HTTP requests):

typescript

export default {
  async fetch(request: Request, env: Env, ctx: ExecutionContext) {
    return new Response("Hello!");
  },
};

Scheduled Handler (cron jobs):

typescript

export default {
  async scheduled(event: ScheduledEvent, env: Env, ctx: ExecutionContext) {
    // Runs on schedule defined in wrangler.toml
    await env.MY_KV.put("last-run", new Date().toISOString());
  },
};

Queue Handler (message processing):

typescript

export default {
  async queue(batch: MessageBatch<any>, env: Env, ctx: ExecutionContext) {
    for (const message of batch.messages) {
      await processMessage(message.body);
      message.ack();
    }
  },
};

Bindings#

Bindings connect your Worker to Cloudflare resources. Configure in wrangler.toml:

KV (Key-Value Storage):

toml

[[kv_namespaces]]
binding = "MY_KV"
id = "your-kv-namespace-id"

typescript

// Usage
await env.MY_KV.put("key", "value");
const value = await env.MY_KV.get("key");

D1 (SQL Database):

toml

[[d1_databases]]
binding = "DB"
database_name = "my-database"
database_id = "your-database-id"

typescript

// Usage
const result = await env.DB.prepare(
  "SELECT * FROM users WHERE id = ?"
).bind(userId).all();

R2 (Object Storage):

toml

[[r2_buckets]]
binding = "MY_BUCKET"
bucket_name = "my-bucket"

typescript

// Usage
await env.MY_BUCKET.put("file.txt", "contents");
const object = await env.MY_BUCKET.get("file.txt");
const text = await object?.text();

Environment Variables (non-secret only):

toml

[vars]
GREETING = "hello"
ENVIRONMENT = "development"

Secrets (sensitive data — never put in wrangler.toml):

bash

# Set via CLI; the value is encrypted server-side and never written to disk locally
wrangler secret put MY_API_KEY

Context (ctx)#

The ctx parameter provides control over request lifecycle:

typescript

export default {
  async fetch(request: Request, env: Env, ctx: ExecutionContext) {
    // Run tasks after response is sent
    ctx.waitUntil(
      env.MY_KV.put("request-count", String(Date.now()))
    );

    // Pass through to origin on exception
    ctx.passThroughOnException();

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

Top-level Environment Access#

Since March 2025, you can import env at the module level instead of passing it through handlers:

typescript

import { env } from "cloudflare:workers";

// Access bindings outside of handlers
const apiClient = new ApiClient({ apiKey: env.API_KEY });

export default {
  async fetch(request: Request): Promise<Response> {
    // env is also available here without the parameter
    const data = await env.MY_KV.get("config");
    return Response.json({ data });
  },
};

This eliminates prop-drilling env through function signatures and enables module-level initialization.

Rapid Development Patterns#

Wrangler Configuration#

Essential wrangler.toml:

toml

name = "my-worker"
main = "src/index.ts"
compatibility_date = "2025-09-01"

# Custom domain
routes = [
  { pattern = "api.example.com/*", zone_name = "example.com" }
]

# Or workers.dev subdomain
workers_dev = true

# Environment variables
[vars]
ENVIRONMENT = "production"

# Bindings
[[kv_namespaces]]
binding = "CACHE"
id = "your-kv-id"

[[d1_databases]]
binding = "DB"
database_name = "production-db"
database_id = "your-db-id"

[[r2_buckets]]
binding = "ASSETS"
bucket_name = "my-assets"

# Cron triggers
[triggers]
crons = ["0 0 * * *"]  # Daily at midnight

Environment Management#

Use environments for staging/production:

toml

[env.staging]
vars = { ENVIRONMENT = "staging" }

[env.staging.d1_databases]
binding = "DB"
database_name = "staging-db"
database_id = "staging-db-id"

[env.production]
vars = { ENVIRONMENT = "production" }

[env.production.d1_databases]
binding = "DB"
database_name = "production-db"
database_id = "production-db-id"

bash

# Deploy to staging
wrangler deploy --env staging

# Deploy to production
wrangler deploy --env production

Common Patterns#

JSON API with Error Handling:

typescript

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

      if (url.pathname === "/api/users" && request.method === "GET") {
        const users = await env.DB.prepare("SELECT * FROM users").all();
        return Response.json(users.results);
      }

      if (url.pathname === "/api/users" && request.method === "POST") {
        const body = await request.json();
        await env.DB.prepare(
          "INSERT INTO users (name, email) VALUES (?, ?)"
        ).bind(body.name, body.email).run();
        return Response.json({ success: true }, { status: 201 });
      }

      return Response.json({ error: "Not found" }, { status: 404 });
    } catch (error) {
      return Response.json(
        { error: error.message },
        { status: 500 }
      );
    }
  },
};

Authentication Middleware:

typescript

async function authenticate(request: Request, env: Env): Promise<string | null> {
  const authHeader = request.headers.get("Authorization");
  if (!authHeader?.startsWith("Bearer ")) {
    return null;
  }

  const token = authHeader.substring(7);
  const userId = await env.SESSIONS.get(token);
  return userId;
}

export default {
  async fetch(request: Request, env: Env): Promise<Response> {
    const userId = await authenticate(request, env);

    if (!userId) {
      return Response.json({ error: "Unauthorized" }, { status: 401 });
    }

    // Proceed with authenticated request
    return Response.json({ userId });
  },
};

CORS Headers:

typescript

const corsHeaders = {
  "Access-Control-Allow-Origin": "*",
  "Access-Control-Allow-Methods": "GET, POST, PUT, DELETE, OPTIONS",
  "Access-Control-Allow-Headers": "Content-Type, Authorization",
};

export default {
  async fetch(request: Request): Promise<Response> {
    if (request.method === "OPTIONS") {
      return new Response(null, { headers: corsHeaders });
    }

    const response = await handleRequest(request);

    // Add CORS headers to response
    Object.entries(corsHeaders).forEach(([key, value]) => {
      response.headers.set(key, value);
    });

    return response;
  },
};

Static Assets (Full-Stack Apps)#

Serve static files alongside your Worker code:

toml

[assets]
directory = "./public"
binding = "ASSETS"

typescript

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

    // API routes
    if (url.pathname.startsWith("/api/")) {
      return handleAPI(request, env);
    }

    // Serve static assets via the ASSETS binding
    return env.ASSETS.fetch(request);
  },
};

Testing#

Using Vitest:

typescript

import { env, createExecutionContext } from "cloudflare:test";
import { describe, it, expect } from "vitest";
import worker from "./index";

describe("Worker", () => {
  it("responds with JSON", async () => {
    const request = new Request("http://example.com/api/hello");
    const ctx = createExecutionContext();
    const response = await worker.fetch(request, env, ctx);

    expect(response.status).toBe(200);
    expect(await response.json()).toEqual({ message: "Hello!" });
  });
});

Framework Integration#

Workers supports major frameworks with adapters:

Next.js - Full App Router and Pages Router support
Remix / React Router - Native Cloudflare adapter
Astro - Server-side rendering on Workers
SvelteKit - Cloudflare adapter available
Hono - Lightweight web framework built for Workers
tRPC - Type-safe APIs with full Workers support

Example with Hono:

typescript

import { Hono } from "hono";

const app = new Hono();

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

export default app;

Advanced Topics#

For detailed information on advanced features, see the reference files:

Complete Bindings Guide: references/bindings-complete-guide.md - All binding types (D1, KV, R2, Durable Objects, Queues, Workers AI, Vectorize, Workflows, Containers, Secrets Store, Pipelines, AutoRAG)
Deployment & CI/CD: references/wrangler-and-deployment.md - Wrangler v4 migration, commands, GitHub Actions, GitLab CI/CD, gradual rollouts, remote bindings
Development Best Practices: references/development-patterns.md - Testing, debugging, error handling, performance, top-level env access patterns
Advanced Features: references/advanced-features.md - Containers, Workflows, MCP servers, Workers for Platforms, WebSockets, Node.js compat, streaming
Observability: references/observability.md - Logging (tail, Logpush, Workers Logs), metrics, traces, debugging