api-gateway-testing

API gateway testing skill covering rate limiting validation, request routing, authentication proxy testing, load balancing verification, circuit breaker testing, and gateway configuration validation for Kong, Envoy, and AWS API Gateway.

Ir para a instalação

Repositório: PramodDutta/qaskills
Última atividade na origem: 17 de março de 2026 às 07:05
Idioma detectado do SKILL.md: inglês
Estrelas: 201
Forks: 21

Opções de instalação

Por padrão, está selecionado o prompt que primeiro revisa a origem. Você pode mudar para um comando direto ou baixar uma cópia local.

Revise os arquivos de origem

Leia o SKILL.md e os arquivos complementares exibidos pelo SkillsMP antes de decidir se vai instalar.

Voltar à pesquisa Voltar à instalação

name	API Gateway Testing
description	API gateway testing skill covering rate limiting validation, request routing, authentication proxy testing, load balancing verification, circuit breaker testing, and gateway configuration validation for Kong, Envoy, and AWS API Gateway.
version	1.0.0
author	thetestingacademy
license	MIT
tags	["api-gateway","rate-limiting","routing","kong","envoy","aws-api-gateway","load-balancing","circuit-breaker"]
testingTypes	["integration","api","performance"]
frameworks	["playwright","jest","pytest"]
languages	["typescript","javascript","python"]
domains	["api","backend","devops"]
agents	["claude-code","cursor","github-copilot","windsurf","codex","aider","continue","cline","zed","bolt","gemini-cli","amp"]

API Gateway Testing Skill

You are an expert QA automation engineer specializing in API gateway testing. When the user asks you to write, review, or debug tests for API gateways including rate limiting, routing, authentication proxying, circuit breakers, and gateway configuration validation, follow these detailed instructions.

Core Principles

Test the gateway, not the backend -- Isolate gateway behavior from upstream services. Mock backends when testing routing, rate limiting, and transformation rules.
Deterministic rate limit validation -- Rate limit tests must account for clock skew, sliding windows, and reset timing. Always verify both the allow and deny states.
Contract-first verification -- Every gateway route should be tested against its OpenAPI specification or route configuration contract.
Failure mode coverage -- Gateways are critical infrastructure. Test circuit breaker tripping, failover routing, timeout handling, and retry behavior explicitly.
Security boundary testing -- The gateway is the first line of defense. Verify authentication enforcement, CORS policies, header injection prevention, and TLS termination.
Environment parity -- Gateway configurations often differ between dev, staging, and production. Test configuration loading and environment-specific overrides.
Observability validation -- Verify that the gateway emits correct access logs, metrics, and tracing headers for every request path.

Project Structure

Always organize API gateway testing projects with this structure:

tests/
  gateway/
    routing/
      path-routing.test.ts
      header-routing.test.ts
      method-routing.test.ts
    rate-limiting/
      fixed-window.test.ts
      sliding-window.test.ts
      per-client.test.ts
    auth/
      jwt-validation.test.ts
      api-key.test.ts
      oauth-proxy.test.ts
    circuit-breaker/
      trip-threshold.test.ts
      recovery.test.ts
    transformation/
      request-transform.test.ts
      response-transform.test.ts
    cors/
      cors-policy.test.ts
    load-balancing/
      round-robin.test.ts
      weighted.test.ts
  fixtures/
    mock-backend/
      server.ts
      routes.ts
    gateway-config/
      kong.yaml
      envoy.yaml
  utils/
    gateway-client.ts
    rate-limit-helpers.ts
    jwt-helpers.ts
    mock-server.ts
  config/
    test-config.ts
jest.config.ts

Gateway Client Utility

// utils/gateway-client.ts
import axios, { AxiosInstance, AxiosResponse }  ;

  {
  : ;
  ?: ;
  ?: ;
  ?: ;
}

   {
   : ;

  () {
     : <, > = {};
     (config.) headers[] = config.;
     (config.) headers[] = ;

    . = axios.({
      : config.,
      : config. || ,
      headers,
      :  , 
    });
  }

   (: , ?: <, >): <> {
     ..(path, { headers });
  }

   (: , ?: , ?: <, >): <> {
     ..(path, data, { headers });
  }

   (: , ?: , ?: <, >): <> {
     ..(path, data, { headers });
  }

   (: , ?: <, >): <> {
     ..(path, { headers });
  }

   (
    :  | ,
    : ,
    : ,
    ?: <, >
  ): <[]> {
     requests = .({ : count }, 
      method ===  ? .(path, headers) : .(path, {}, headers)
    );
     .(requests);
  }
}

import { describe, it, expect, beforeAll, afterAll } from '@jest/globals';
import { GatewayClient } from '../utils/gateway-client';

describe('Fixed Window Rate Limiting', () => {
  const gateway = new GatewayClient({
    baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
    apiKey: 'test-api-key-rate-limit',
  });

  it('should allow requests within the rate limit', async () => {
    const RATE_LIMIT = 100; // 100 requests per minute

    // Send requests within limit
    const responses = await gateway.sendConcurrent('GET', '/api/v1/products', RATE_LIMIT - 10);

    const successResponses = responses.filter((r) => r.status === 200);
    expect(successResponses.length).toBe(RATE_LIMIT - 10);

    // Verify rate limit headers
    const lastResponse = responses[responses.length - 1];
    expect(lastResponse.headers['x-ratelimit-limit']).toBe('100');
    expect(parseInt(lastResponse.headers['x-ratelimit-remaining'])).toBeGreaterThan(0);
  });

  it('should reject requests exceeding the rate limit with 429', async () => {
    const RATE_LIMIT = 100;

    // Exhaust the rate limit
    const responses = await gateway.sendConcurrent('GET', '/api/v1/products', RATE_LIMIT + 20);

    const rejectedResponses = responses.filter((r) => r.status === 429);
    expect(rejectedResponses.length).toBeGreaterThan(0);

    // Verify 429 response body
    const rejected = rejectedResponses[0];
    expect(rejected.data).toHaveProperty('message');
    expect(rejected.data.message).toContain('rate limit');

    // Verify Retry-After header
    expect(rejected.headers['retry-after']).toBeDefined();
    const retryAfter = parseInt(rejected.headers['retry-after']);
    expect(retryAfter).toBeGreaterThan(0);
    expect(retryAfter).toBeLessThanOrEqual(60);
  });

  it('should apply rate limits per client independently', async () => {
    const clientA = new GatewayClient({
      baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
      apiKey: 'client-a-key',
    });

    const clientB = new GatewayClient({
      baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
      apiKey: 'client-b-key',
    });

    // Client A exhausts their limit
    await clientA.sendConcurrent('GET', '/api/v1/products', 110);

    // Client B should still be able to make requests
    const responseB = await clientB.get('/api/v1/products');
    expect(responseB.status).toBe(200);
  });

  it('should reset rate limit after the window expires', async () => {
    const RATE_LIMIT = 100;

    // Exhaust the rate limit
    await gateway.sendConcurrent('GET', '/api/v1/products', RATE_LIMIT + 10);

    // Verify we are rate limited
    const blockedResponse = await gateway.get('/api/v1/products');
    expect(blockedResponse.status).toBe(429);

    // Wait for window reset (check Retry-After header)
    const retryAfter = parseInt(blockedResponse.headers['retry-after']) || 60;
    await new Promise((resolve) => setTimeout(resolve, retryAfter * 1000));

    // Should be allowed again
    const recoveredResponse = await gateway.get('/api/v1/products');
    expect(recoveredResponse.status).toBe(200);
  }, 120000);
});

describe('Sliding Window Rate Limiting', () => {
  const gateway = new GatewayClient({
    baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
    apiKey: 'sliding-window-key',
  });

  it('should enforce sliding window limits across time boundaries', async () => {
    // Send 50 requests in first half of the window
    const firstBatch = await gateway.sendConcurrent('GET', '/api/v1/data', 50);
    expect(firstBatch.filter((r) => r.status === 200).length).toBe(50);

    // Wait half the window
    await new Promise((resolve) => setTimeout(resolve, 30000));

    // Send another 60 requests -- should hit limit since window slides
    const secondBatch = await gateway.sendConcurrent('GET', '/api/v1/data', 60);
    const rejected = secondBatch.filter((r) => r.status === 429);
    expect(rejected.length).toBeGreaterThan(0);
  }, 120000);
});

import { describe, it, expect } from '@jest/globals';
import { GatewayClient } from '../utils/gateway-client';

describe('Request Routing', () => {
  const gateway = new GatewayClient({
    baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
  });

  it('should route requests based on path prefix', async () => {
    const usersResponse = await gateway.get('/api/v1/users');
    expect(usersResponse.status).toBe(200);
    expect(usersResponse.headers['x-upstream']).toBe('users-service');

    const productsResponse = await gateway.get('/api/v1/products');
    expect(productsResponse.status).toBe(200);
    expect(productsResponse.headers['x-upstream']).toBe('products-service');

    const ordersResponse = await gateway.get('/api/v1/orders');
    expect(ordersResponse.status).toBe(200);
    expect(ordersResponse.headers['x-upstream']).toBe('orders-service');
  });

  it('should route based on request headers', async () => {
    const v1Response = await gateway.get('/api/products', {
      'Accept-Version': 'v1',
    });
    expect(v1Response.headers['x-api-version']).toBe('v1');

    const v2Response = await gateway.get('/api/products', {
      'Accept-Version': 'v2',
    });
    expect(v2Response.headers['x-api-version']).toBe('v2');
  });

  it('should route based on HTTP method', async () => {
    const getResponse = await gateway.get('/api/v1/resources');
    expect(getResponse.status).toBe(200);

    const postResponse = await gateway.post('/api/v1/resources', { name: 'test' });
    expect(postResponse.status).toBe(201);

    const deleteResponse = await gateway.delete('/api/v1/resources/123');
    expect(deleteResponse.status).toBe(204);
  });

  it('should return 404 for unmatched routes', async () => {
    const response = await gateway.get('/api/v1/nonexistent-endpoint');
    expect(response.status).toBe(404);
    expect(response.data).toHaveProperty('error');
  });

  it('should handle path parameter extraction correctly', async () => {
    const response = await gateway.get('/api/v1/users/12345/orders');
    expect(response.status).toBe(200);
    // Backend should receive the extracted path parameters
    expect(response.data.userId).toBe('12345');
  });

  it('should preserve query parameters through routing', async () => {
    const response = await gateway.get('/api/v1/products?page=2&limit=10&sort=price');
    expect(response.status).toBe(200);
    expect(response.data.pagination).toEqual(
      expect.objectContaining({
        page: 2,
        limit: 10,
      })
    );
  });
});

import { describe, it, expect } from '@jest/globals';
import { GatewayClient } from '../utils/gateway-client';
import { generateJWT, generateExpiredJWT } from '../utils/jwt-helpers';

describe('Authentication Proxy', () => {
  it('should reject requests without authentication', async () => {
    const gateway = new GatewayClient({
      baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
    });

    const response = await gateway.get('/api/v1/protected/resource');
    expect(response.status).toBe(401);
    expect(response.data.error).toContain('authentication required');
  });

  it('should accept valid JWT tokens', async () => {
    const token = generateJWT({ sub: 'user-123', role: 'admin' });
    const gateway = new GatewayClient({
      baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
      jwtToken: token,
    });

    const response = await gateway.get('/api/v1/protected/resource');
    expect(response.status).toBe(200);
    // Gateway should forward user identity to upstream
    expect(response.headers['x-user-id']).toBe('user-123');
  });

  it('should reject expired JWT tokens', async () => {
    const expiredToken = generateExpiredJWT({ sub: 'user-123' });
    const gateway = new GatewayClient({
      baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
      jwtToken: expiredToken,
    });

    const response = await gateway.get('/api/v1/protected/resource');
    expect(response.status).toBe(401);
    expect(response.data.error).toContain('token expired');
  });

  it('should reject JWT with invalid signature', async () => {
    const tampered = generateJWT({ sub: 'user-123' }, 'wrong-secret-key');
    const gateway = new GatewayClient({
      baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
      jwtToken: tampered,
    });

    const response = await gateway.get('/api/v1/protected/resource');
    expect(response.status).toBe(401);
  });

  it('should validate API key authentication', async () => {
    const validKeyGateway = new GatewayClient({
      baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
      apiKey: 'valid-api-key-12345',
    });

    const response = await validKeyGateway.get('/api/v1/data');
    expect(response.status).toBe(200);

    const invalidKeyGateway = new GatewayClient({
      baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
      apiKey: 'invalid-key',
    });

    const rejectedResponse = await invalidKeyGateway.get('/api/v1/data');
    expect(rejectedResponse.status).toBe(403);
  });

  it('should enforce role-based access control', async () => {
    const adminToken = generateJWT({ sub: 'admin-1', role: 'admin' });
    const viewerToken = generateJWT({ sub: 'viewer-1', role: 'viewer' });

    const adminGateway = new GatewayClient({
      baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
      jwtToken: adminToken,
    });
    const viewerGateway = new GatewayClient({
      baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
      jwtToken: viewerToken,
    });

    // Admin can delete
    const adminDelete = await adminGateway.delete('/api/v1/resources/123');
    expect(adminDelete.status).toBe(204);

    // Viewer cannot delete
    const viewerDelete = await viewerGateway.delete('/api/v1/resources/456');
    expect(viewerDelete.status).toBe(403);
  });
});

import { describe, it, expect, beforeEach } from '@jest/globals';
import { GatewayClient } from '../utils/gateway-client';
import { MockServer } from '../utils/mock-server';

describe('Circuit Breaker', () => {
  const gateway = new GatewayClient({
    baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
    apiKey: 'test-key',
  });
  let mockBackend: MockServer;

  beforeEach(async () => {
    mockBackend = new MockServer(3001);
    await mockBackend.start();
  });

  afterEach(async () => {
    await mockBackend.stop();
  });

  it('should trip circuit breaker after consecutive failures', async () => {
    // Configure mock to return 500 errors
    mockBackend.setResponseCode(500);

    const FAILURE_THRESHOLD = 5;

    // Send enough requests to trip the circuit breaker
    for (let i = 0; i < FAILURE_THRESHOLD + 2; i++) {
      await gateway.get('/api/v1/unstable-service');
    }

    // Circuit should now be open -- gateway returns 503
    const response = await gateway.get('/api/v1/unstable-service');
    expect(response.status).toBe(503);
    expect(response.data.error).toContain('circuit breaker');
  });

  it('should recover after the circuit breaker timeout', async () => {
    // Trip the circuit
    mockBackend.setResponseCode(500);
    for (let i = 0; i < 10; i++) {
      await gateway.get('/api/v1/unstable-service');
    }

    // Verify circuit is open
    const openResponse = await gateway.get('/api/v1/unstable-service');
    expect(openResponse.status).toBe(503);

    // Fix the backend
    mockBackend.setResponseCode(200);

    // Wait for circuit breaker timeout (half-open state)
    await new Promise((resolve) => setTimeout(resolve, 30000));

    // Next request should be allowed through (half-open test)
    const recoveryResponse = await gateway.get('/api/v1/unstable-service');
    expect(recoveryResponse.status).toBe(200);
  }, 60000);

  it('should not trip circuit breaker for client errors (4xx)', async () => {
    mockBackend.setResponseCode(400);

    // Send many 400 errors
    for (let i = 0; i < 20; i++) {
      await gateway.get('/api/v1/unstable-service');
    }

    // Circuit should remain closed -- 400s are client errors, not backend failures
    mockBackend.setResponseCode(200);
    const response = await gateway.get('/api/v1/unstable-service');
    expect(response.status).toBe(200);
  });
});

describe('Request/Response Transformation', () => {
  const gateway = new GatewayClient({
    baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
    apiKey: 'test-key',
  });

  it('should add required headers to upstream requests', async () => {
    const response = await gateway.get('/api/v1/echo-headers');

    // Gateway should add correlation ID
    expect(response.data.receivedHeaders['x-correlation-id']).toBeDefined();
    expect(response.data.receivedHeaders['x-correlation-id']).toMatch(
      /^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$/
    );

    // Gateway should add request timestamp
    expect(response.data.receivedHeaders['x-request-timestamp']).toBeDefined();
  });

  it('should strip sensitive headers from responses', async () => {
    const response = await gateway.get('/api/v1/products');

    // Internal headers should be stripped
    expect(response.headers['x-internal-service-id']).toBeUndefined();
    expect(response.headers['x-debug-trace']).toBeUndefined();
    expect(response.headers['server']).toBeUndefined();
  });

  it('should transform response body format based on Accept header', async () => {
    const jsonResponse = await gateway.get('/api/v1/products', {
      Accept: 'application/json',
    });
    expect(jsonResponse.headers['content-type']).toContain('application/json');

    const xmlResponse = await gateway.get('/api/v1/products', {
      Accept: 'application/xml',
    });
    expect(xmlResponse.headers['content-type']).toContain('application/xml');
  });
});

describe('CORS Policy Enforcement', () => {
  const gateway = new GatewayClient({
    baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
  });

  it('should allow requests from whitelisted origins', async () => {
    const response = await gateway.get('/api/v1/products', {
      Origin: 'https://app.example.com',
    });

    expect(response.headers['access-control-allow-origin']).toBe('https://app.example.com');
    expect(response.status).toBe(200);
  });

  it('should reject requests from non-whitelisted origins', async () => {
    const response = await gateway.get('/api/v1/products', {
      Origin: 'https://malicious-site.com',
    });

    expect(response.headers['access-control-allow-origin']).toBeUndefined();
  });

  it('should handle preflight OPTIONS requests correctly', async () => {
    const response = await gateway.client.options('/api/v1/products', {
      headers: {
        Origin: 'https://app.example.com',
        'Access-Control-Request-Method': 'POST',
        'Access-Control-Request-Headers': 'Content-Type, Authorization',
      },
    });

    expect(response.status).toBe(204);
    expect(response.headers['access-control-allow-methods']).toContain('POST');
    expect(response.headers['access-control-allow-headers']).toContain('Authorization');
    expect(response.headers['access-control-max-age']).toBeDefined();
  });

  it('should not allow disallowed HTTP methods', async () => {
    const response = await gateway.client.options('/api/v1/products', {
      headers: {
        Origin: 'https://app.example.com',
        'Access-Control-Request-Method': 'PATCH',
      },
    });

    const allowedMethods = response.headers['access-control-allow-methods'] || '';
    expect(allowedMethods).not.toContain('PATCH');
  });
});

# test_gateway.py
import pytest
import requests
import time
from concurrent.futures import ThreadPoolExecutor, as_completed

GATEWAY_URL = "http://localhost:8000"

class TestRateLimiting:
    """Test rate limiting at the API gateway."""

    def test_allows_requests_within_limit(self):
        """Should allow requests within the configured rate limit."""
        headers = {"X-API-Key": "test-key"}
        responses = []

        for _ in range(50):
            resp = requests.get(f"{GATEWAY_URL}/api/v1/products", headers=headers)
            responses.append(resp)

        success_count = sum(1 for r in responses if r.status_code == 200)
        assert success_count == 50

    def test_rejects_requests_exceeding_limit(self):
        """Should return 429 when rate limit is exceeded."""
        headers = {"X-API-Key": "burst-test-key"}

        def make_request():
            return requests.get(f"{GATEWAY_URL}/api/v1/products", headers=headers)

        with ThreadPoolExecutor(max_workers=20) as executor:
            futures = [executor.submit(make_request) for _ in range(150)]
            responses = [f.result() for f in as_completed(futures)]

        status_codes = [r.status_code for r in responses]
        assert 429 in status_codes, "Expected at least one 429 response"

    def test_rate_limit_headers_present(self):
        """Should include rate limit headers in responses."""
        headers = {"X-API-Key": "header-test-key"}
        resp = requests.get(f"{GATEWAY_URL}/api/v1/products", headers=headers)

        assert "X-RateLimit-Limit" in resp.headers
        assert "X-RateLimit-Remaining" in resp.headers
        assert "X-RateLimit-Reset" in resp.headers


class TestCircuitBreaker:
    """Test circuit breaker behavior at the gateway."""

    def test_circuit_opens_after_failures(self):
        """Should open circuit after consecutive backend failures."""
        headers = {"X-API-Key": "circuit-test-key"}

        # Trigger backend failures
        for _ in range(10):
            requests.get(f"{GATEWAY_URL}/api/v1/failing-service", headers=headers)

        # Circuit should be open
        resp = requests.get(f"{GATEWAY_URL}/api/v1/failing-service", headers=headers)
        assert resp.status_code == 503
        assert "circuit" in resp.json().get("error", "").lower()

    def test_circuit_recovers(self):
        """Should allow requests after circuit breaker timeout."""
        headers = {"X-API-Key": "recovery-test-key"}

        # Trip the circuit
        for _ in range(10):
            requests.get(f"{GATEWAY_URL}/api/v1/flaky-service", headers=headers)

        # Wait for recovery window
        time.sleep(30)

        resp = requests.get(f"{GATEWAY_URL}/api/v1/flaky-service", headers=headers)
        assert resp.status_code in [200, 503]  # Half-open state may succeed or fail


class TestHealthChecks:
    """Test gateway health check endpoints."""

    def test_gateway_health_endpoint(self):
        resp = requests.get(f"{GATEWAY_URL}/health")
        assert resp.status_code == 200
        data = resp.json()
        assert data["status"] == "healthy"

    def test_gateway_readiness_check(self):
        resp = requests.get(f"{GATEWAY_URL}/ready")
        assert resp.status_code == 200
        data = resp.json()
        assert "upstreams" in data
        for upstream in data["upstreams"]:
            assert upstream["status"] in ["healthy", "degraded"]

    def test_gateway_liveness_check(self):
        resp = requests.get(f"{GATEWAY_URL}/live")
        assert resp.status_code == 200

describe('Gateway Failover', () => {
  const gateway = new GatewayClient({
    baseURL: process.env.GATEWAY_URL || 'http://localhost:8000',
    apiKey: 'failover-key',
  });

  it('should failover to secondary upstream when primary is down', async () => {
    // Simulate primary being down (mock server stopped)
    const response = await gateway.get('/api/v1/critical-service');

    // Gateway should route to secondary upstream
    expect(response.status).toBe(200);
    expect(response.headers['x-served-by']).toMatch(/secondary|backup/);
  });

  it('should return cached response when all upstreams are unavailable', async () => {
    // First request populates cache
    const initialResponse = await gateway.get('/api/v1/cacheable-data');
    expect(initialResponse.status).toBe(200);

    // All upstreams down -- should serve stale cache
    const cachedResponse = await gateway.get('/api/v1/cacheable-data');
    expect(cachedResponse.status).toBe(200);
    expect(cachedResponse.headers['x-cache-status']).toBe('STALE');
  });

  it('should respect timeout configuration for slow upstreams', async () => {
    const startTime = Date.now();
    const response = await gateway.get('/api/v1/slow-service');
    const duration = Date.now() - startTime;

    // Gateway should timeout before the slow service responds
    expect(response.status).toBe(504);
    expect(duration).toBeLessThan(10000); // Should timeout within configured limit
  });
});