当验证码解决在凌晨 3 点失败时,"Error solving captcha" 等纯文本日志没有帮助。包含任务 ID、验证码类型、解决时间和错误代码的结构化 JSON 日志可让您过滤、搜索并准确地发出错误警报。
为什么要结构化日志记录
| 纯文本 | 结构化 JSON |
|---|---|
Captcha solved in 12.3s |
{"event":"captcha_solved","task_id":"abc123","type":"recaptcha_v2","solve_time_ms":12300} |
| 难以解析 | 机器可读 |
| 仅限 Grep 搜索 | 按任意字段过滤 |
| 没有相关性 | 任务ID链接提交→轮询→注入 |
Python:结构日志
import structlog
import time
structlog.configure(
processors=[
structlog.processors.TimeStamper(fmt="iso"),
structlog.processors.add_log_level,
structlog.processors.JSONRenderer(),
],
logger_factory=structlog.PrintLoggerFactory(),
)
log = structlog.get_logger()
记录求解生命周期
import requests
API_KEY = "YOUR_API_KEY"
def solve_captcha(captcha_type, sitekey, page_url, proxy=None):
solve_log = log.bind(
captcha_type=captcha_type,
site_url=page_url,
sitekey=sitekey[:12] + "...",
)
# Submit
start = time.time()
solve_log.info("captcha_submit_start")
resp = requests.post("https://ocr.captchaai.com/in.php", data={
"key": API_KEY,
"method": "userrecaptcha",
"googlekey": sitekey,
"pageurl": page_url,
"json": "1",
}).json()
if resp["status"] != 1:
solve_log.error("captcha_submit_failed", error=resp["request"])
return None
task_id = resp["request"]
submit_ms = int((time.time() - start) * 1000)
solve_log = solve_log.bind(task_id=task_id)
solve_log.info("captcha_submitted", submit_ms=submit_ms)
# Poll
for attempt in range(24):
time.sleep(5)
result = requests.get("https://ocr.captchaai.com/res.php", params={
"key": API_KEY, "action": "get", "id": task_id, "json": "1"
}).json()
if result["status"] == 1:
solve_ms = int((time.time() - start) * 1000)
solve_log.info(
"captcha_solved",
solve_time_ms=solve_ms,
poll_attempts=attempt + 1,
token_length=len(result["request"]),
)
return result["request"]
if result["request"] != "CAPCHA_NOT_READY":
solve_log.error(
"captcha_solve_failed",
error=result["request"],
poll_attempts=attempt + 1,
)
return None
solve_log.warning("captcha_solve_timeout", poll_attempts=24)
return None
输出:
{"event":"captcha_submit_start","captcha_type":"recaptcha_v2","site_url":"https://example.com","sitekey":"6Le-wvkSAAAA...","timestamp":"2025-07-15T10:30:00Z","level":"info"}
{"event":"captcha_submitted","task_id":"71845302","submit_ms":245,"timestamp":"2025-07-15T10:30:00Z","level":"info"}
{"event":"captcha_solved","task_id":"71845302","solve_time_ms":18230,"poll_attempts":4,"token_length":580,"timestamp":"2025-07-15T10:30:18Z","level":"info"}
Node.js:皮诺
const pino = require('pino');
const log = pino({
level: 'info',
timestamp: pino.stdTimeFunctions.isoTime,
});
记录求解生命周期
const axios = require('axios');
const API_KEY = 'YOUR_API_KEY';
async function solveCaptcha(captchaType, sitekey, pageUrl) {
const taskLog = log.child({
captchaType,
siteUrl: pageUrl,
sitekey: sitekey.substring(0, 12) + '...',
});
const start = Date.now();
taskLog.info('captcha_submit_start');
const submit = await axios.post('https://ocr.captchaai.com/in.php', null, {
params: {
key: API_KEY, method: 'userrecaptcha',
googlekey: sitekey, pageurl: pageUrl, json: 1,
},
});
if (submit.data.status !== 1) {
taskLog.error({ error: submit.data.request }, 'captcha_submit_failed');
return null;
}
const taskId = submit.data.request;
const boundLog = taskLog.child({ taskId });
boundLog.info({ submitMs: Date.now() - start }, 'captcha_submitted');
for (let attempt = 1; attempt <= 24; attempt++) {
await new Promise(r => setTimeout(r, 5000));
const poll = await axios.get('https://ocr.captchaai.com/res.php', {
params: { key: API_KEY, action: 'get', id: taskId, json: 1 },
});
if (poll.data.status === 1) {
boundLog.info({
solveTimeMs: Date.now() - start,
pollAttempts: attempt,
tokenLength: poll.data.request.length,
}, 'captcha_solved');
return poll.data.request;
}
if (poll.data.request !== 'CAPCHA_NOT_READY') {
boundLog.error({ error: poll.data.request, pollAttempts: attempt }, 'captcha_solve_failed');
return null;
}
}
boundLog.warn({ pollAttempts: 24 }, 'captcha_solve_timeout');
return null;
}
日志字段参考
| 场地 | 类型 | 描述 |
|---|---|---|
event |
细绳 | 活动名称:captcha_submitted、captcha_solved等 |
task_id |
细绳 | CaptchaAI 用于关联的任务 ID |
captcha_type |
细绳 | recaptcha_v2、turnstile、image 等 |
site_url |
细绳 | 目标页面网址 |
solve_time_ms |
整数 | 从提交到解决的总时间 |
poll_attempts |
整数 | 提出的投票请求数量 |
error |
细绳 | 来自 CaptchaAI 的错误代码 |
token_length |
整数 | 返回token的长度 |
过滤和警报
查找最近一小时内的所有故障
# With jq
cat captcha.log | jq 'select(.level == "error" and .event == "captcha_solve_failed")'
高故障率警报
# Count errors vs successes in a rolling window
from collections import deque
class ErrorRateMonitor:
def __init__(self, window_size=100, threshold=0.2):
self.results = deque(maxlen=window_size)
self.threshold = threshold
def record(self, success):
self.results.append(success)
if len(self.results) >= 50:
error_rate = 1 - sum(self.results) / len(self.results)
if error_rate > self.threshold:
log.warning(
"captcha_error_rate_high",
error_rate=round(error_rate, 3),
window=len(self.results),
)
故障排除
| 问题 | 原因 | 处理方式 |
|---|---|---|
| 日志太详细 | 记录每次投票尝试 | 仅记录提交、已解决和失败事件 |
| 无法关联事件 | 缺少任务 ID | 尽早将 task_id 与 log.bind() 或 log.child() 绑定 |
| 日志不可搜索 | 纯文本格式 | 使用 structlog 或 pino 切换到 JSON |
| 日志中的敏感数据 | 记录完整的 API 密钥 | 切勿记录 API 密钥;截断站点密钥 |
常问问题
我应该记录验证码令牌吗?
记录令牌长度,而不是完整令牌。令牌可以超过 500 个字符,并增加不必要的日志量。
CAPCHA_NOT_READY 的日志级别是什么?
根本不记录它——这是轮询期间所期望的。仅记录最终结果。
使用 CaptchaAI 构建可观察的验证码工作流程
获取您的 API 密钥:验证码网站。