- Add MsgConnectionRefusedException to distinguish ECONNREFUSED from other connection errors; openConnection() now throws it when connect() returns ECONNREFUSED. - MsgClient::call() adds retry_on_refused/retry_interval parameters (defaults false/1.0s), matching Python pymsg.MsgClient.call(retry_on_refused=True). - Retry loop: only retries on ECONNREFUSED; other exceptions (timeout, connection reset, business errors) propagate immediately. Throttled logging: first failure prints immediately, subsequent logs every 5 seconds. - DbC::call_msg() enables retry_on_refused=true (db is a hard dependency; gateway blocks waiting for db on startup instead of crashing). - LogC::_send() explicitly passes retry_on_refused=false (fail-safe; never block the background log thread when db is unavailable). - captcha/smtp auxiliary services keep default retry_on_refused=false (fail-fast; blocking in request threads would exhaust the thread pool). - Update README with pymsg protocol documentation, downstream retry strategy table, and v0.6.1 changelog entry.
250 lines
8.1 KiB
C++
250 lines
8.1 KiB
C++
#pragma once
|
||
|
||
/** LogC — gateway 统一日志客户端(C++ 版,对应各 Python 服务的 log_client.py)
|
||
*
|
||
* 设计要点(与 Python 版对齐):
|
||
* - 通过 pymsg TCP 调用 dbv2 execute 路由,逐条 INSERT 到 Logs 表
|
||
* - 异步非阻塞:内存队列 + 后台线程批量推送,不拖慢业务
|
||
* - fail-safe:db 服务不可用时静默丢弃(仅 stderr 提示),绝不影响业务
|
||
* - 缓冲批量:队列满 50 条或超时 2 秒时自动 flush
|
||
*
|
||
* 用法:
|
||
* #include "LogC.hpp"
|
||
*
|
||
* qing::LogC::instance().info("gateway", "启动完成");
|
||
* qing::LogC::instance().error("gateway", "数据库连接失败",
|
||
* std::nullopt, R"({"errno":2002})");
|
||
*
|
||
* // 进程退出前:
|
||
* qing::LogC::instance().shutdown(); // 停止后台线程并 flush 剩余日志(幂等)
|
||
*/
|
||
|
||
#include <atomic>
|
||
#include <chrono>
|
||
#include <condition_variable>
|
||
#include <cstdio>
|
||
#include <ctime>
|
||
#include <deque>
|
||
#include <iostream>
|
||
#include <mutex>
|
||
#include <optional>
|
||
#include <string>
|
||
#include <thread>
|
||
#include <vector>
|
||
|
||
#include <boost/json.hpp>
|
||
|
||
#include "Msg.hpp"
|
||
|
||
namespace qing {
|
||
|
||
namespace json = boost::json;
|
||
|
||
class LogC {
|
||
public:
|
||
struct Entry {
|
||
std::string service;
|
||
std::string level;
|
||
std::string message;
|
||
std::optional<std::string> trace_id;
|
||
std::optional<std::string> extra; // 已序列化的 JSON 字符串(存 Logs.extra TEXT)
|
||
};
|
||
|
||
static LogC& instance() {
|
||
static LogC inst;
|
||
return inst;
|
||
}
|
||
|
||
// ---- 公开接口 --------------------------------------------------------
|
||
|
||
// 推送一条日志(非阻塞,立即返回)
|
||
void push(const std::string& service, const std::string& level,
|
||
const std::string& message,
|
||
std::optional<std::string> trace_id = std::nullopt,
|
||
std::optional<std::string> extra = std::nullopt) {
|
||
{
|
||
std::lock_guard<std::mutex> lock(_mtx);
|
||
_queue.push_back(Entry{service, level, message,
|
||
std::move(trace_id), std::move(extra)});
|
||
}
|
||
_cv.notify_one();
|
||
}
|
||
|
||
void debug(const std::string& service, const std::string& message,
|
||
std::optional<std::string> trace_id = std::nullopt,
|
||
std::optional<std::string> extra = std::nullopt) {
|
||
push(service, "DEBUG", message, std::move(trace_id), std::move(extra));
|
||
}
|
||
void info(const std::string& service, const std::string& message,
|
||
std::optional<std::string> trace_id = std::nullopt,
|
||
std::optional<std::string> extra = std::nullopt) {
|
||
push(service, "INFO", message, std::move(trace_id), std::move(extra));
|
||
}
|
||
void warning(const std::string& service, const std::string& message,
|
||
std::optional<std::string> trace_id = std::nullopt,
|
||
std::optional<std::string> extra = std::nullopt) {
|
||
push(service, "WARNING", message, std::move(trace_id), std::move(extra));
|
||
}
|
||
void error(const std::string& service, const std::string& message,
|
||
std::optional<std::string> trace_id = std::nullopt,
|
||
std::optional<std::string> extra = std::nullopt) {
|
||
push(service, "ERROR", message, std::move(trace_id), std::move(extra));
|
||
}
|
||
void critical(const std::string& service, const std::string& message,
|
||
std::optional<std::string> trace_id = std::nullopt,
|
||
std::optional<std::string> extra = std::nullopt) {
|
||
push(service, "CRITICAL", message, std::move(trace_id), std::move(extra));
|
||
}
|
||
|
||
// 手动 flush:通知后台线程立即发送当前积攒的批次,并短暂等待其完成
|
||
void flush() {
|
||
_force_flush.store(true);
|
||
_cv.notify_all();
|
||
for (int i = 0; i < 200 && _force_flush.load(); ++i) {
|
||
std::this_thread::sleep_for(std::chrono::milliseconds(10));
|
||
}
|
||
}
|
||
|
||
// 停止后台线程并 flush 剩余日志(进程退出前调用,幂等)
|
||
void shutdown() {
|
||
{
|
||
std::lock_guard<std::mutex> lock(_mtx);
|
||
if (_stop) return;
|
||
_stop = true;
|
||
}
|
||
_cv.notify_all();
|
||
if (_thread.joinable()) _thread.join();
|
||
}
|
||
|
||
// ---- 配置(构造后、首次推送前可改)------------------------------------
|
||
void set_endpoint(const std::string& addr, int port) {
|
||
_db_addr = addr;
|
||
_db_port = port;
|
||
}
|
||
|
||
private:
|
||
LogC() : _thread(&LogC::_flush_loop, this) {}
|
||
~LogC() { shutdown(); }
|
||
|
||
LogC(const LogC&) = delete;
|
||
LogC& operator=(const LogC&) = delete;
|
||
|
||
// 后台线程:积攒到 batch_size 或超时 flush_interval 时发送
|
||
void _flush_loop() {
|
||
std::vector<Entry> batch;
|
||
auto last_flush = std::chrono::steady_clock::now();
|
||
|
||
for (;;) {
|
||
{
|
||
std::unique_lock<std::mutex> lock(_mtx);
|
||
// 最多 0.5s 轮询一次(对齐 Python 版 chk_timeout 上限)
|
||
_cv.wait_for(lock, std::chrono::milliseconds(500),
|
||
[&] { return _stop || !_queue.empty(); });
|
||
while (!_queue.empty() && batch.size() < _BATCH_SIZE) {
|
||
batch.push_back(std::move(_queue.front()));
|
||
_queue.pop_front();
|
||
}
|
||
}
|
||
|
||
const auto now = std::chrono::steady_clock::now();
|
||
const bool due = std::chrono::duration<double>(now - last_flush).count()
|
||
>= _FLUSH_INTERVAL;
|
||
if (!batch.empty() && (batch.size() >= _BATCH_SIZE || due
|
||
|| _stop || _force_flush.load())) {
|
||
_send(batch);
|
||
batch.clear();
|
||
last_flush = now;
|
||
_force_flush.store(false);
|
||
}
|
||
|
||
if (_stop) {
|
||
// 退出前 flush 队列剩余
|
||
{
|
||
std::unique_lock<std::mutex> lock(_mtx);
|
||
while (!_queue.empty()) {
|
||
batch.push_back(std::move(_queue.front()));
|
||
_queue.pop_front();
|
||
}
|
||
}
|
||
if (!batch.empty()) _send(batch);
|
||
_force_flush.store(false);
|
||
return;
|
||
}
|
||
}
|
||
}
|
||
|
||
// 通过 pymsg 使用 execute 路由逐条 INSERT 到 dbv2 Logs 表
|
||
void _send(const std::vector<Entry>& batch) {
|
||
const std::string now = _utc_now_iso(); // 批次共享同一时间戳(对齐 Python 版)
|
||
try {
|
||
pmc::net::MsgClient client(_db_addr, _db_port, "gateway-log");
|
||
for (const auto& e : batch) {
|
||
json::object values;
|
||
values["ts"] = now;
|
||
values["service"] = e.service;
|
||
values["level"] = e.level;
|
||
values["message"] = e.message;
|
||
values["trace_id"] = e.trace_id ? json::value(*e.trace_id)
|
||
: json::value(nullptr);
|
||
values["extra"] = e.extra ? json::value(*e.extra)
|
||
: json::value(nullptr);
|
||
|
||
json::object payload;
|
||
payload["_route"] = "execute";
|
||
payload["action"] = "insert";
|
||
payload["table"] = "Logs";
|
||
payload["values"] = std::move(values);
|
||
|
||
const std::string resp = client.call(
|
||
json::serialize(payload), _TIMEOUT,
|
||
/*retry_on_refused=*/false); // 日志是 fail-safe,db 不可用时静默丢弃,绝不阻塞重试
|
||
auto obj = json::parse(resp).as_object();
|
||
if (!obj.at("success").as_bool()) {
|
||
std::string err = "unknown";
|
||
if (auto* ev = obj.if_contains("error")) {
|
||
if (ev->is_string()) err = ev->as_string().c_str();
|
||
}
|
||
std::cerr << "[LogC] INSERT 失败: " << err << std::endl;
|
||
}
|
||
}
|
||
} catch (const std::exception& e) {
|
||
// fail-safe:db 不可用时静默丢弃,绝不影响业务
|
||
std::cerr << "[LogC] 发送失败: " << e.what() << std::endl;
|
||
}
|
||
}
|
||
|
||
// UTC ISO8601 时间戳(对齐 Python datetime.now(timezone.utc).isoformat())
|
||
// 形如 2026-07-29T08:30:00.123456+00:00,字典序即时间序,便于 db 按 ts 清理
|
||
static std::string _utc_now_iso() {
|
||
const auto now = std::chrono::system_clock::now();
|
||
const auto t = std::chrono::system_clock::to_time_t(now);
|
||
const auto us = std::chrono::duration_cast<std::chrono::microseconds>(
|
||
now.time_since_epoch()).count() % 1000000;
|
||
std::tm tm{};
|
||
::gmtime_r(&t, &tm);
|
||
char buf[40];
|
||
std::snprintf(buf, sizeof buf,
|
||
"%04d-%02d-%02dT%02d:%02d:%02d.%06d+00:00",
|
||
tm.tm_year + 1900, tm.tm_mon + 1, tm.tm_mday,
|
||
tm.tm_hour, tm.tm_min, tm.tm_sec, static_cast<int>(us));
|
||
return buf;
|
||
}
|
||
|
||
// ---- 配置项 -----------------------------------------------------------
|
||
std::string _db_addr = "127.0.0.1";
|
||
int _db_port = 8005;
|
||
static constexpr size_t _BATCH_SIZE = 50; // 队列满 50 条时 flush
|
||
static constexpr double _FLUSH_INTERVAL = 2.0; // 超时 2 秒强制 flush
|
||
static constexpr double _TIMEOUT = 5.0; // socket 超时(秒)
|
||
|
||
// ---- 状态 -------------------------------------------------------------
|
||
std::deque<Entry> _queue;
|
||
std::mutex _mtx;
|
||
std::condition_variable _cv;
|
||
bool _stop = false;
|
||
std::atomic<bool> _force_flush{false};
|
||
std::thread _thread;
|
||
};
|
||
|
||
} // namespace qing
|