AGENT SKILL

apollo-rate-limits

Implement Apollo.io rate limiting and backoff. Use when handling rate limits, implementing retry logic, or optimizing API request throughput. Trigger with phrases like "apollo rate limit", "apollo 429", "apollo throttling", "apollo backoff", "apollo request limits".

License 许可
MIT
最近更新
2026-11-15
Stars
★ 2.0k
Forks
233

01 前言 Front Matter

name
apollo-rate-limits
author
Jeremy Longshore <jeremy@intentsolutions.io>
license
MIT
version
1.0.0
description
Implement Apollo.io rate limiting and backoff. Use when handling rate limits, implementing retry logic, or optimizing API request throughput. Trigger with phrases like "apollo rate limit", "apollo 429", "apollo throttling", "apollo backoff", "apollo request limits".
allowed-tools
Read, Grep, Bash(curl:*)

03 SKILL.md 正文

Apollo Rate Limits

Overview

Implement robust rate limiting and backoff strategies for Apollo.io API to maximize throughput while avoiding 429 errors.

Apollo Rate Limits

Endpoint Category Rate Limit Window Burst Limit
People Search 100/min 1 minute 10/sec
Person Enrichment 100/min 1 minute 10/sec
Organization Enrichment 100/min 1 minute 10/sec
Sequences/Campaigns 50/min 1 minute 5/sec
Bulk Operations 10/min 1 minute 2/sec
General API 100/min 1 minute 10/sec

Rate Limit Headers

# Check current rate limit status
curl -I -X POST "https://api.apollo.io/v1/people/search" \
  -H "Content-Type: application/json" \
  -d '{"api_key": "'$APOLLO_API_KEY'", "per_page": 1}'

# Response headers:
# X-RateLimit-Limit: 100
# X-RateLimit-Remaining: 95
# X-RateLimit-Reset: 1640000000
# Retry-After: 60  (only when rate limited)

Implementation: Rate Limiter Class

// src/lib/apollo/rate-limiter.ts
interface RateLimiterConfig {
  maxRequests: number;
  windowMs: number;
  minSpacingMs: number;
}

class RateLimiter {
  private queue: Array<{
    resolve: (value: void) => void;
    reject: (error: Error) => void;
  }> = [];
  private requestTimestamps: number[] = [];
  private lastRequestTime = 0;
  private processing = false;

  constructor(private config: RateLimiterConfig) {}

  async acquire(): Promise<void> {
    return new Promise((resolve, reject) => {
      this.queue.push({ resolve, reject });
      this.processQueue();
    });
  }

  private async processQueue() {
    if (this.processing || this.queue.length === 0) return;
    this.processing = true;

    while (this.queue.length > 0) {
      // Clean old timestamps outside window
      const now = Date.now();
      this.requestTimestamps = this.requestTimestamps.filter(
        (ts) => now - ts < this.config.windowMs
      );

      // Check if we're at capacity
      if (this.requestTimestamps.length >= this.config.maxRequests) {
        const oldestTs = this.requestTimestamps[0];
        const waitTime = this.config.windowMs - (now - oldestTs) + 100;
        await this.wait(waitTime);
        continue;
      }

      // Enforce minimum spacing
      const timeSinceLastRequest = now - this.lastRequestTime;
      if (timeSinceLastRequest < this.config.minSpacingMs) {
        await this.wait(this.config.minSpacingMs - timeSinceLastRequest);
      }

      // Process next request
      const item = this.queue.shift()!;
      this.requestTimestamps.push(Date.now());
      this.lastRequestTime = Date.now();
      item.resolve();
    }

    this.processing = false;
  }

  private wait(ms: number): Promise<void> {
    return new Promise((resolve) => setTimeout(resolve, ms));
  }
}

// Create rate limiter for Apollo
export const apolloRateLimiter = new RateLimiter({
  maxRequests: 90, // Leave buffer below 100
  windowMs: 60000,
  minSpacingMs: 100, // 100ms between requests
});

Implementation: Exponential Backoff

// src/lib/apollo/backoff.ts
interface BackoffConfig {
  initialDelayMs: number;
  maxDelayMs: number;
  maxRetries: number;
  multiplier: number;
  jitter: boolean;
}

const defaultConfig: BackoffConfig = {
  initialDelayMs: 1000,
  maxDelayMs: 60000,
  maxRetries: 5,
  multiplier: 2,
  jitter: true,
};

export async function withBackoff<T>(
  fn: () => Promise<T>,
  config: Partial<BackoffConfig> = {}
): Promise<T> {
  const cfg = { ...defaultConfig, ...config };
  let lastError: Error;
  let delay = cfg.initialDelayMs;

  for (let attempt = 0; attempt <= cfg.maxRetries; attempt++) {
    try {
      await apolloRateLimiter.acquire();
      return await fn();
    } catch (error: any) {
      lastError = error;

      // Check if retryable
      const status = error.response?.status;
      if (status === 401 || status === 403 || status === 422) {
        throw error; // Don't retry auth/validation errors
      }

      if (attempt === cfg.maxRetries) {
        break;
      }

      // Get delay from Retry-After header or calculate
      const retryAfter = error.response?.headers?.['retry-after'];
      if (retryAfter) {
        delay = parseInt(retryAfter) * 1000;
      }

      // Add jitter to prevent thundering herd
      const jitter = cfg.jitter ? Math.random() * 1000 : 0;
      const actualDelay = Math.min(delay + jitter, cfg.maxDelayMs);

      console.log(`Retry ${attempt + 1}/${cfg.maxRetries} after ${actualDelay}ms`);
      await new Promise((r) => setTimeout(r, actualDelay));

      delay *= cfg.multiplier;
    }
  }

  throw lastError!;
}

Implementation: Request Queue

// src/lib/apollo/request-queue.ts
import PQueue from 'p-queue';

// Concurrency-limited queue
export const apolloQueue = new PQueue({
  concurrency: 5, // Max 5 concurrent requests
  interval: 1000, // Per second
  intervalCap: 10, // Max 10 per interval
});

// Usage
async function batchSearchPeople(domains: string[]): Promise<Person[]> {
  const results = await Promise.all(
    domains.map((domain) =>
      apolloQueue.add(() =>
        withBackoff(() => apollo.searchPeople({ q_organization_domains: [domain] }))
      )
    )
  );

  return results.flat().map((r) => r?.people || []).flat();
}

Usage Patterns

Pattern 1: Simple Rate-Limited Request

import { withBackoff } from './backoff';

const people = await withBackoff(() =>
  apollo.searchPeople({
    q_organization_domains: ['stripe.com'],
    per_page: 100,
  })
);

Pattern 2: Batch Processing with Queue

import { apolloQueue } from './request-queue';

async function enrichCompanies(domains: string[]) {
  const results = [];

  for (const domain of domains) {
    const result = await apolloQueue.add(
      () => withBackoff(() => apollo.enrichOrganization(domain)),
      { priority: 1 } // Lower priority
    );
    results.push(result);
  }

  return results;
}

Pattern 3: Priority Queue for Interactive vs Background

// High priority for user-facing requests
async function interactiveSearch(query: string) {
  return apolloQueue.add(
    () => withBackoff(() => apollo.searchPeople({ q_keywords: query })),
    { priority: 0 } // Highest priority
  );
}

// Low priority for background sync
async function backgroundSync(contacts: string[]) {
  return Promise.all(
    contacts.map((id) =>
      apolloQueue.add(
        () => withBackoff(() => apollo.getContact(id)),
        { priority: 10 } // Low priority
      )
    )
  );
}

Monitoring Rate Limit Usage

// src/lib/apollo/rate-monitor.ts
class RateLimitMonitor {
  private requests: Array<{ timestamp: number; remaining: number }> = [];

  recordRequest(remaining: number) {
    this.requests.push({
      timestamp: Date.now(),
      remaining,
    });

    // Keep only last 5 minutes
    const cutoff = Date.now() - 5 * 60 * 1000;
    this.requests = this.requests.filter((r) => r.timestamp > cutoff);
  }

  getStats() {
    const lastMinute = this.requests.filter(
      (r) => r.timestamp > Date.now() - 60000
    );

    return {
      requestsLastMinute: lastMinute.length,
      currentRemaining: lastMinute[lastMinute.length - 1]?.remaining ?? 100,
      utilizationPercent: (lastMinute.length / 100) * 100,
      isNearLimit: lastMinute.length > 80,
    };
  }
}

export const rateLimitMonitor = new RateLimitMonitor();

Output

  • Rate limiter class with token bucket algorithm
  • Exponential backoff with jitter
  • Request queue with concurrency control
  • Priority-based request scheduling
  • Rate limit monitoring and alerts

Error Handling

Scenario Strategy
429 response Use Retry-After header
Burst limit hit Add minimum spacing
Sustained limit Queue with concurrency
Network timeout Exponential backoff

Resources

Next Steps

Proceed to apollo-security-basics for API security best practices.

04 同仓库更多 Skills

Dicklesworthstone
Dicklesworthstone/pi_agent_rust

theme-factory

Toolkit for styling artifacts with a theme. These artifacts can be slides, docs, reportings, HTML landing pages, etc. There are 10 pre-set themes with colors/fonts that you can apply to any artifact that has been creating, or can generate a new theme on-the-fly.

视觉设计
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

web-artifacts-builder

Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn/ui). Use for complex artifacts requiring state management, routing, or shadcn/ui components - not for simple single-file HTML/JSX artifacts.

前端开发
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

webapp-testing

Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.

测试与质量前端开发
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

neuropixels-analysis

Neuropixels neural recording analysis. Load SpikeGLX/OpenEphys data, preprocess, motion correction, Kilosort4 spike sorting, quality metrics, Allen/IBL curation, AI-assisted visual analysis, for Neuropixels 1.0/2.0 extracellular electrophysiology. Use when working with neural recordings, spike sorting, extracellular electrophysiology, or when the user mentions Neuropixels, SpikeGLX, Open Ephys, Kilosort, quality metrics, or unit curation.

★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

skill-creator

Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.

Skill与智能体开发
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

notebooklm

Use this skill to query your Google NotebookLM notebooks directly from Claude Code for source-grounded, citation-backed answers from Gemini. Browser automation, library management, persistent auth. Drastically reduced hallucinations through document-only responses.

Skill与智能体开发
★ 2.0k