api-design-patterns（API 设计模式）

API Design Patterns

Resource Naming

• Use plural nouns: /users, /orders, /products
• Nest for relationships: /users/{id}/orders
• Max nesting depth: 2 levels. Beyond that, use query params or top-level resources
• Use kebab-case: /user-profiles, not /userProfiles
• Never put verbs in URLs: /users/{id}/activate is wrong, use POST /users/{id}/activation

HTTP Methods

Return Location header on POST with the URL of the created resource.

Status Codes

200 OK              - Successful read/update
201 Created         - Successful creation
204 No Content      - Successful delete
400 Bad Request     - Validation error (include field-level errors)
401 Unauthorized    - Missing or invalid authentication
403 Forbidden       - Authenticated but not authorized
404 Not Found       - Resource does not exist
409 Conflict        - State conflict (duplicate, version mismatch)
422 Unprocessable   - Semantically invalid (valid JSON, bad values)
429 Too Many Reqs   - Rate limited (include Retry-After header)
500 Internal Error  - Unhandled server error (never expose stack traces)

Error Response Format

{
  "error": {
    "code": "VALIDATION_ERROR",
    "message": "Request validation failed",
    "details": [
      { "field": "email", "message": "Must be a valid email address" },
      { "field": "age", "message": "Must be at least 18" }
    ]
  }
}

Use consistent error codes across the API. Document every code in your API reference.

Cursor-Based Pagination (preferred)

GET /users?limit=20&cursor=eyJpZCI6MTAwfQ

Response:
{
  "data": [...],
  "pagination": {
    "next_cursor": "eyJpZCI6MTIwfQ",
    "has_more": true
  }
}

Use cursor pagination for large or frequently changing datasets. Encode cursors as opaque base64 strings. Never expose raw IDs in cursors.

Offset-Based Pagination (simple cases only)

GET /users?page=3&per_page=20

Response:
{
  "data": [...],
  "pagination": {
    "page": 3,
    "per_page": 20,
    "total": 245,
    "total_pages": 13
  }
}

Only use offset pagination when total count is cheap and dataset is small.

Filtering and Sorting

GET /orders?status=pending&created_after=2025-01-01&sort=-created_at,+total
GET /products?category=electronics&price_min=100&price_max=500
GET /users?search=john&fields=id,name,email

Use field selection (fields param) to reduce payload size. Prefix sort fields with - for descending.

Versioning

Prefer URL path versioning for simplicity:

/api/v1/users
/api/v2/users

Rules:

• Never break v1 once published. Add fields, never remove them.
• New required fields = new version
• Deprecate old versions with Sunset header and 6-month notice
• Support at most 2 active versions simultaneously

Request/Response Headers

Content-Type: application/json
Accept: application/json
Authorization: Bearer <token>
X-Request-Id: <uuid>          # For tracing
X-RateLimit-Limit: 100        # Requests per window
X-RateLimit-Remaining: 47     # Remaining in window
X-RateLimit-Reset: 1700000000 # Window reset Unix timestamp
Retry-After: 30               # Seconds until rate limit resets

Always return X-Request-Id in responses for debugging.

OpenAPI Spec Guidelines

• Write spec first, then implement (spec-driven development)
• Use $ref for shared schemas: $ref: '#/components/schemas/User'
• Define examples for every endpoint
• Use oneOf/anyOf for polymorphic responses
• Generate client SDKs from the spec, never hand-write them
• Validate requests against the spec in middleware

paths:
  /users/{id}:
    get:
      operationId: getUser
      parameters:
        - name: id
          in: path
          required: true
          schema:
            type: string
            format: uuid
      responses:
        '200':
          description: User found
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/User'
        '404':
          $ref: '#/components/responses/NotFound'

Rate Limiting Strategy

• Apply per-user, per-endpoint limits
• Use sliding window algorithm (not fixed window)
• Return 429 with Retry-After header
• Exempt health check and auth endpoints from rate limits
• Log rate-limited requests for abuse detection