Rate Limits

Esta página explica os limites de requisições (rate limits) da API Vexy Bank e como implementar estratégias eficientes para respeitá-los.

Limites por endpoint

Tabela de Rate Limits

Endpoint	Limite	Janela de Tempo	Tipo
`POST /api/auth`	5 requests	1 minuto	Por IP + API Key
`POST /api/v1/pix/in/qrcode`	100 requests	1 minuto	Por API Key
`POST /api/v1/pix/out/pixkey`	50 requests	1 minuto	Por API Key
`GET /api/v1/webhook`	30 requests	1 minuto	Por API Key
`POST /api/v1/webhook`	10 requests	1 minuto	Por API Key
`DELETE /api/v1/webhook/:id`	10 requests	1 minuto	Por API Key

Headers de Rate Limit

A API retorna headers informativos sobre seus limites:

X-RateLimit-Limit: 100
X-RateLimit-Remaining: 85
X-RateLimit-Reset: 1640995260
X-RateLimit-Window: 60

Header	Descrição
`X-RateLimit-Limit`	Limite máximo para o endpoint
`X-RateLimit-Remaining`	Requisições restantes na janela atual
`X-RateLimit-Reset`	Timestamp quando o limite será resetado
`X-RateLimit-Window`	Tamanho da janela em segundos

Quando o limite é excedido

Resposta HTTP 429

{
  "success": false,
  "error": {
    "code": "RATE_LIMIT_EXCEEDED",
    "message": "Rate limit exceeded. Try again in 42 seconds.",
    "details": {
      "limit": 100,
      "window": 60,
      "retryAfter": 42
    }
  },
  "timestamp": "2024-01-20T10:30:00.000Z"
}

Headers Adicionais

HTTP/1.1 429 Too Many Requests
Retry-After: 42
X-RateLimit-Limit: 100
X-RateLimit-Remaining: 0
X-RateLimit-Reset: 1640995302

Implementação de rate limiting

Cliente com Rate Limiting Inteligente

class RateLimitedAPIClient {
  constructor(apiKey, apiSecret, baseUrl) {
    this.auth = new VexyBankAuth(apiKey, apiSecret, baseUrl)
    this.rateLimits = new Map()
    this.requestQueue = new Map()
  }

  async makeRequest(method, endpoint, data = null) {
    // Verificar se precisa aguardar
    await this.waitIfRateLimited(endpoint)
    
    try {
      const response = await this.auth.makeAuthenticatedRequest(method, endpoint, data)
      
      // Atualizar informações de rate limit
      this.updateRateLimitInfo(endpoint, response.headers)
      
      return response
    } catch (error) {
      if (error.status === 429) {
        await this.handleRateLimit(endpoint, error)
        // Tentar novamente após aguardar
        return await this.makeRequest(method, endpoint, data)
      }
      throw error
    }
  }

  async waitIfRateLimited(endpoint) {
    const limitInfo = this.rateLimits.get(endpoint)
    
    if (!limitInfo) return
    
    const now = Math.floor(Date.now() / 1000)
    
    // Se ainda estamos na janela de rate limit e sem requests restantes
    if (limitInfo.resetTime > now && limitInfo.remaining <= 0) {
      const waitTime = (limitInfo.resetTime - now) * 1000
      console.log(`Rate limit ativo para ${endpoint}. Aguardando ${waitTime}ms...`)
      await new Promise(resolve => setTimeout(resolve, waitTime))
    }
  }

  updateRateLimitInfo(endpoint, headers) {
    if (headers['x-ratelimit-limit']) {
      this.rateLimits.set(endpoint, {
        limit: parseInt(headers['x-ratelimit-limit']),
        remaining: parseInt(headers['x-ratelimit-remaining']),
        resetTime: parseInt(headers['x-ratelimit-reset']),
        window: parseInt(headers['x-ratelimit-window'])
      })
    }
  }

  async handleRateLimit(endpoint, error) {
    const retryAfter = error.response?.headers['retry-after'] || 60
    console.log(`Rate limit excedido para ${endpoint}. Aguardando ${retryAfter} segundos...`)
    
    // Atualizar informações de rate limit
    this.rateLimits.set(endpoint, {
      limit: 0,
      remaining: 0,
      resetTime: Math.floor(Date.now() / 1000) + parseInt(retryAfter),
      window: parseInt(retryAfter)
    })
    
    await new Promise(resolve => setTimeout(resolve, parseInt(retryAfter) * 1000))
  }

  getRateLimitStatus(endpoint) {
    return this.rateLimits.get(endpoint) || null
  }

  getAllRateLimitStatus() {
    const status = {}
    for (const [endpoint, info] of this.rateLimits.entries()) {
      const now = Math.floor(Date.now() / 1000)
      status[endpoint] = {
        ...info,
        isActive: info.resetTime > now,
        timeToReset: Math.max(0, info.resetTime - now)
      }
    }
    return status
  }
}

Sistema de Fila de Requisições

class RequestQueue {
  constructor(rateLimitedClient) {
    this.client = rateLimitedClient
    this.queues = new Map()
    this.processing = new Map()
  }

  async enqueue(endpoint, method, data = null, priority = 'normal') {
    if (!this.queues.has(endpoint)) {
      this.queues.set(endpoint, [])
    }

    const request = {
      id: this.generateRequestId(),
      method,
      endpoint,
      data,
      priority,
      timestamp: Date.now(),
      resolve: null,
      reject: null
    }

    return new Promise((resolve, reject) => {
      request.resolve = resolve
      request.reject = reject
      
      const queue = this.queues.get(endpoint)
      
      if (priority === 'high') {
        queue.unshift(request)
      } else {
        queue.push(request)
      }
      
      this.processQueue(endpoint)
    })
  }

  async processQueue(endpoint) {
    if (this.processing.get(endpoint)) {
      return // Já processando
    }

    this.processing.set(endpoint, true)

    try {
      const queue = this.queues.get(endpoint)
      
      while (queue.length > 0) {
        const request = queue.shift()
        
        try {
          const result = await this.client.makeRequest(
            request.method,
            request.endpoint,
            request.data
          )
          request.resolve(result)
        } catch (error) {
          request.reject(error)
        }
        
        // Pequena pausa entre requisições
        await new Promise(resolve => setTimeout(resolve, 100))
      }
    } finally {
      this.processing.set(endpoint, false)
    }
  }

  generateRequestId() {
    return `req_${Date.now()}_${Math.random().toString(36).substr(2, 9)}`
  }

  getQueueStatus() {
    const status = {}
    for (const [endpoint, queue] of this.queues.entries()) {
      status[endpoint] = {
        pending: queue.length,
        processing: this.processing.get(endpoint) || false
      }
    }
    return status
  }
}

// Uso
const client = new RateLimitedAPIClient(API_KEY, API_SECRET, BASE_URL)
const queue = new RequestQueue(client)

// Requisições com prioridade
await queue.enqueue('/api/v1/pix/in/qrcode', 'POST', paymentData, 'high')
await queue.enqueue('/api/v1/pix/out/pixkey', 'POST', transferData, 'normal')

Otimização de performance

Batch Operations

class BatchProcessor {
  constructor(client, options = {}) {
    this.client = client
    this.batchSize = options.batchSize || 10
    this.batchDelay = options.batchDelay || 1000
    this.batches = new Map()
  }

  async addToBatch(endpoint, operation) {
    if (!this.batches.has(endpoint)) {
      this.batches.set(endpoint, [])
      // Processar batch após delay
      setTimeout(() => this.processBatch(endpoint), this.batchDelay)
    }

    const batch = this.batches.get(endpoint)
    
    return new Promise((resolve, reject) => {
      batch.push({ operation, resolve, reject })
      
      // Processar se batch está cheio
      if (batch.length >= this.batchSize) {
        this.processBatch(endpoint)
      }
    })
  }

  async processBatch(endpoint) {
    const batch = this.batches.get(endpoint)
    if (!batch || batch.length === 0) return

    this.batches.set(endpoint, []) // Limpar batch

    console.log(`Processando batch de ${batch.length} operações para ${endpoint}`)

    // Processar com rate limiting
    for (const { operation, resolve, reject } of batch) {
      try {
        const result = await this.client.makeRequest(
          operation.method,
          operation.endpoint,
          operation.data
        )
        resolve(result)
      } catch (error) {
        reject(error)
      }
      
      // Pausa entre operações do batch
      await new Promise(resolve => setTimeout(resolve, 100))
    }
  }
}

// Uso
const batchProcessor = new BatchProcessor(client, {
  batchSize: 5,
  batchDelay: 2000
})

// Adicionar operações ao batch
const promises = []
for (let i = 0; i < 20; i++) {
  promises.push(
    batchProcessor.addToBatch('/api/v1/pix/in/qrcode', {
      method: 'POST',
      endpoint: '/api/v1/pix/in/qrcode',
      data: paymentData[i]
    })
  )
}

const results = await Promise.all(promises)

Cache Inteligente

class CachedAPIClient {
  constructor(client, options = {}) {
    this.client = client
    this.cache = new Map()
    this.defaultTTL = options.defaultTTL || 300000 // 5 minutos
    this.cacheable = options.cacheable || ['GET']
  }

  async makeRequest(method, endpoint, data = null, options = {}) {
    const cacheKey = this.getCacheKey(method, endpoint, data)
    const useCache = this.cacheable.includes(method) && !options.skipCache

    // Tentar cache primeiro
    if (useCache) {
      const cached = this.getFromCache(cacheKey)
      if (cached) {
        console.log(`Cache hit para ${endpoint}`)
        return cached
      }
    }

    try {
      const result = await this.client.makeRequest(method, endpoint, data)
      
      // Cachear resultado
      if (useCache) {
        this.setCache(cacheKey, result, options.ttl || this.defaultTTL)
      }
      
      return result
    } catch (error) {
      // Em caso de rate limit, tentar cache mesmo que expirado
      if (error.status === 429 && useCache) {
        const stale = this.getFromCache(cacheKey, true)
        if (stale) {
          console.log(`Usando cache expirado devido ao rate limit para ${endpoint}`)
          return stale
        }
      }
      throw error
    }
  }

  getCacheKey(method, endpoint, data) {
    const dataHash = data ? this.hashObject(data) : ''
    return `${method}:${endpoint}:${dataHash}`
  }

  getFromCache(key, allowStale = false) {
    const item = this.cache.get(key)
    if (!item) return null
    
    if (!allowStale && Date.now() > item.expiresAt) {
      this.cache.delete(key)
      return null
    }
    
    return item.data
  }

  setCache(key, data, ttl) {
    this.cache.set(key, {
      data,
      expiresAt: Date.now() + ttl,
      createdAt: Date.now()
    })
  }

  hashObject(obj) {
    return require('crypto')
      .createHash('md5')
      .update(JSON.stringify(obj))
      .digest('hex')
      .substring(0, 8)
  }

  clearCache() {
    this.cache.clear()
  }

  getCacheStats() {
    let expired = 0
    let active = 0
    const now = Date.now()
    
    for (const [key, item] of this.cache.entries()) {
      if (now > item.expiresAt) {
        expired++
      } else {
        active++
      }
    }
    
    return { total: this.cache.size, active, expired }
  }
}

Monitoramento de rate limits

Dashboard de Métricas

class RateLimitMonitor {
  constructor() {
    this.metrics = {
      requests: new Map(),
      rateLimitHits: new Map(),
      averageWaitTime: new Map()
    }
  }

  recordRequest(endpoint, timestamp = Date.now()) {
    if (!this.metrics.requests.has(endpoint)) {
      this.metrics.requests.set(endpoint, [])
    }
    
    this.metrics.requests.get(endpoint).push(timestamp)
    this.cleanOldMetrics(endpoint)
  }

  recordRateLimitHit(endpoint, waitTime) {
    if (!this.metrics.rateLimitHits.has(endpoint)) {
      this.metrics.rateLimitHits.set(endpoint, [])
    }
    
    this.metrics.rateLimitHits.get(endpoint).push({
      timestamp: Date.now(),
      waitTime
    })
    
    this.updateAverageWaitTime(endpoint, waitTime)
  }

  updateAverageWaitTime(endpoint, waitTime) {
    const current = this.metrics.averageWaitTime.get(endpoint) || { total: 0, count: 0 }
    current.total += waitTime
    current.count += 1
    this.metrics.averageWaitTime.set(endpoint, current)
  }

  getRequestRate(endpoint, windowMs = 60000) {
    const requests = this.metrics.requests.get(endpoint) || []
    const cutoff = Date.now() - windowMs
    const recentRequests = requests.filter(ts => ts > cutoff)
    
    return {
      endpoint,
      requestCount: recentRequests.length,
      requestsPerMinute: (recentRequests.length / windowMs) * 60000,
      rateLimitHits: this.getRateLimitHits(endpoint, windowMs)
    }
  }

  getRateLimitHits(endpoint, windowMs = 60000) {
    const hits = this.metrics.rateLimitHits.get(endpoint) || []
    const cutoff = Date.now() - windowMs
    return hits.filter(hit => hit.timestamp > cutoff).length
  }

  getAverageWaitTime(endpoint) {
    const data = this.metrics.averageWaitTime.get(endpoint)
    if (!data || data.count === 0) return 0
    return data.total / data.count
  }

  cleanOldMetrics(endpoint, maxAge = 3600000) { // 1 hora
    const cutoff = Date.now() - maxAge
    
    // Limpar requisições antigas
    const requests = this.metrics.requests.get(endpoint) || []
    this.metrics.requests.set(endpoint, requests.filter(ts => ts > cutoff))
    
    // Limpar rate limit hits antigos
    const hits = this.metrics.rateLimitHits.get(endpoint) || []
    this.metrics.rateLimitHits.set(endpoint, hits.filter(hit => hit.timestamp > cutoff))
  }

  generateReport() {
    const report = {
      timestamp: new Date().toISOString(),
      endpoints: {}
    }

    for (const endpoint of this.getAllEndpoints()) {
      const rate = this.getRequestRate(endpoint)
      report.endpoints[endpoint] = {
        requestsPerMinute: rate.requestsPerMinute,
        rateLimitHits: rate.rateLimitHits,
        averageWaitTime: this.getAverageWaitTime(endpoint),
        efficiency: this.calculateEfficiency(endpoint)
      }
    }

    return report
  }

  getAllEndpoints() {
    const endpoints = new Set()
    for (const endpoint of this.metrics.requests.keys()) {
      endpoints.add(endpoint)
    }
    for (const endpoint of this.metrics.rateLimitHits.keys()) {
      endpoints.add(endpoint)
    }
    return Array.from(endpoints)
  }

  calculateEfficiency(endpoint) {
    const rate = this.getRequestRate(endpoint)
    if (rate.requestCount === 0) return 100
    
    const efficiency = ((rate.requestCount - rate.rateLimitHits) / rate.requestCount) * 100
    return Math.round(efficiency * 100) / 100
  }
}

// Uso
const monitor = new RateLimitMonitor()

// Integrar com cliente
const originalMakeRequest = client.makeRequest.bind(client)
client.makeRequest = async function(method, endpoint, data) {
  monitor.recordRequest(endpoint)
  
  try {
    return await originalMakeRequest(method, endpoint, data)
  } catch (error) {
    if (error.status === 429) {
      const waitTime = parseInt(error.response?.headers['retry-after'] || 60) * 1000
      monitor.recordRateLimitHit(endpoint, waitTime)
    }
    throw error
  }
}

// Gerar relatório periódico
setInterval(() => {
  const report = monitor.generateReport()
  console.log('Rate Limit Report:', JSON.stringify(report, null, 2))
}, 60000) // A cada minuto

Estratégias avançadas

Distribuição de Carga

class LoadBalancer {
  constructor(clients) {
    this.clients = clients
    this.currentIndex = 0
  }

  getNextClient() {
    // Round-robin simples
    const client = this.clients[this.currentIndex]
    this.currentIndex = (this.currentIndex + 1) % this.clients.length
    return client
  }

  async makeRequest(method, endpoint, data) {
    const errors = []
    
    // Tentar com cada cliente
    for (let attempt = 0; attempt < this.clients.length; attempt++) {
      const client = this.getNextClient()
      
      try {
        return await client.makeRequest(method, endpoint, data)
      } catch (error) {
        errors.push(error)
        
        // Se não for rate limit, parar tentativas
        if (error.status !== 429) {
          throw error
        }
      }
    }
    
    // Todos os clientes com rate limit
    throw new Error('Todos os clientes estão com rate limit ativo')
  }
}

// Uso com múltiplas API Keys
const clients = [
  new RateLimitedAPIClient(API_KEY_1, API_SECRET_1, BASE_URL),
  new RateLimitedAPIClient(API_KEY_2, API_SECRET_2, BASE_URL),
  new RateLimitedAPIClient(API_KEY_3, API_SECRET_3, BASE_URL)
]

const loadBalancer = new LoadBalancer(clients)

Predição de Rate Limits

class RateLimitPredictor {
  constructor(monitor) {
    this.monitor = monitor
  }

  predictRateLimitRisk(endpoint, plannedRequests) {
    const currentRate = this.monitor.getRequestRate(endpoint)
    const limit = this.getKnownLimit(endpoint)
    
    if (!limit) return { risk: 'unknown', confidence: 0 }
    
    const projectedRequests = currentRate.requestCount + plannedRequests
    const utilizationRate = projectedRequests / limit
    
    let risk = 'low'
    let confidence = 0.8
    
    if (utilizationRate > 0.9) {
      risk = 'high'
      confidence = 0.95
    } else if (utilizationRate > 0.7) {
      risk = 'medium'
      confidence = 0.85
    }
    
    return {
      risk,
      confidence,
      utilizationRate,
      recommendedDelay: this.calculateRecommendedDelay(endpoint, utilizationRate)
    }
  }

  getKnownLimit(endpoint) {
    const limits = {
      '/api/auth': 5,
      '/api/v1/pix/in/qrcode': 100,
      '/api/v1/pix/out/pixkey': 50,
      '/api/v1/webhook': 30
    }
    return limits[endpoint]
  }

  calculateRecommendedDelay(endpoint, utilizationRate) {
    if (utilizationRate < 0.7) return 0
    
    const baseDelay = 1000 // 1 segundo
    const riskMultiplier = utilizationRate > 0.9 ? 3 : 2
    
    return Math.round(baseDelay * riskMultiplier * utilizationRate)
  }
}

Próximos passos

Status e códigos

Referência completa de status e códigos

❌ Tratamento de Erros

Guia completo para tratamento de erros

Segurança

Boas práticas de segurança

PIX IN

Comece a receber pagamentos PIX

Introdução

Autenticação

PIX IN - Recebimento

PIX OUT - Envio

Webhooks

Limites por endpoint

Tabela de Rate Limits

Headers de Rate Limit

Quando o limite é excedido

Resposta HTTP 429

Headers Adicionais

Implementação de rate limiting

Cliente com Rate Limiting Inteligente

Sistema de Fila de Requisições

Otimização de performance

Batch Operations

Cache Inteligente

Monitoramento de rate limits

Dashboard de Métricas

Estratégias avançadas

Distribuição de Carga

Predição de Rate Limits

Próximos passos

Status e códigos

❌ Tratamento de Erros

Segurança

PIX IN

Introdução

Autenticação

PIX IN - Recebimento

PIX OUT - Envio

Webhooks

​Limites por endpoint

​Tabela de Rate Limits

​Headers de Rate Limit

​Quando o limite é excedido

​Resposta HTTP 429

​Headers Adicionais

​Implementação de rate limiting

​Cliente com Rate Limiting Inteligente

​Sistema de Fila de Requisições

​Otimização de performance

​Batch Operations

​Cache Inteligente

​Monitoramento de rate limits

​Dashboard de Métricas

​Estratégias avançadas

​Distribuição de Carga

​Predição de Rate Limits

​Próximos passos

Status e códigos

❌ Tratamento de Erros

Segurança

PIX IN

Limites por endpoint

Tabela de Rate Limits

Headers de Rate Limit

Quando o limite é excedido

Resposta HTTP 429

Headers Adicionais

Implementação de rate limiting

Cliente com Rate Limiting Inteligente

Sistema de Fila de Requisições

Otimização de performance

Batch Operations

Cache Inteligente

Monitoramento de rate limits

Dashboard de Métricas

Estratégias avançadas

Distribuição de Carga

Predição de Rate Limits

Próximos passos