157 lines
5.9 KiB
Python
157 lines
5.9 KiB
Python
"""
|
||
统一日志客户端 — 各服务共用。
|
||
|
||
设计要点：
|
||
- 使用 urllib.request（标准库，不引入 requests）
|
||
- 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，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：
|
||
from log_client import logs
|
||
|
||
logs.info("dnfs", "节点新增成功 id=42")
|
||
logs.error("dnfs", "数据库连接失败", extra={"errno": 2002})
|
||
logs.push("gateway", "WARNING", "登录失败次数超限",
|
||
trace_id="req-abc", extra={"ip": "1.2.3.4"})
|
||
"""
|
||
|
||
import json
|
||
import threading
|
||
import queue
|
||
import time
|
||
import urllib.request
|
||
import urllib.error
|
||
import atexit
|
||
|
||
# ── 配置 ──────────────────────────────────────────────────────────────
|
||
|
||
DB_LOG_URL = 'http://127.0.0.1:8005/log/ingest'
|
||
|
||
_BATCH_SIZE = 50 # 队列满 50 条时 flush
|
||
_FLUSH_INTERVAL = 2.0 # 超时 2 秒强制 flush
|
||
_TIMEOUT = 5 # HTTP 请求超时（秒）
|
||
|
||
|
||
class LogClient:
|
||
"""异步批量日志推送客户端。"""
|
||
|
||
def __init__(self, url=DB_LOG_URL, batch_size=_BATCH_SIZE,
|
||
flush_interval=_FLUSH_INTERVAL):
|
||
self._url = url
|
||
self._batch_size = batch_size
|
||
self._flush_interval = flush_interval
|
||
self._queue = queue.Queue()
|
||
self._stop = threading.Event()
|
||
self._thread = threading.Thread(target=self._flush_loop,
|
||
daemon=True, name='log-client')
|
||
self._thread.start()
|
||
atexit.register(self._shutdown)
|
||
|
||
# ── 公开接口 ────────────────────────────────────────────────────────
|
||
|
||
def push(self, service, level, message, trace_id=None, extra=None):
|
||
"""推送一条日志（非阻塞，立即返回）。"""
|
||
self._queue.put({
|
||
'service': service,
|
||
'level': level,
|
||
'message': message,
|
||
'trace_id': trace_id,
|
||
'extra': extra,
|
||
})
|
||
|
||
def debug(self, service, message, **kw):
|
||
self.push(service, 'DEBUG', message, **kw)
|
||
|
||
def info(self, service, message, **kw):
|
||
self.push(service, 'INFO', message, **kw)
|
||
|
||
def warning(self, service, message, **kw):
|
||
self.push(service, 'WARNING', message, **kw)
|
||
|
||
def error(self, service, message, **kw):
|
||
self.push(service, 'ERROR', message, **kw)
|
||
|
||
def critical(self, service, message, **kw):
|
||
self.push(service, 'CRITICAL', message, **kw)
|
||
|
||
def flush(self):
|
||
"""手动 flush 队列中剩余的日志。"""
|
||
batch = []
|
||
while True:
|
||
try:
|
||
batch.append(self._queue.get_nowait())
|
||
if len(batch) >= self._batch_size:
|
||
self._send(batch)
|
||
batch = []
|
||
except queue.Empty:
|
||
break
|
||
if batch:
|
||
self._send(batch)
|
||
|
||
# ── 内部实现 ────────────────────────────────────────────────────────
|
||
|
||
def _flush_loop(self):
|
||
"""后台线程：积攒到 batch_size 或超时 flush_interval 时发送。"""
|
||
batch = []
|
||
last_flush = time.monotonic()
|
||
|
||
while not self._stop.is_set():
|
||
timeout = max(0, self._flush_interval - (time.monotonic() - last_flush))
|
||
try:
|
||
item = self._queue.get(timeout=timeout)
|
||
batch.append(item)
|
||
# 继续非阻塞地取，直到满 batch_size
|
||
while len(batch) < self._batch_size:
|
||
try:
|
||
batch.append(self._queue.get_nowait())
|
||
except queue.Empty:
|
||
break
|
||
except queue.Empty:
|
||
# 【核心修复 1】：如果没有新日志，必须重置上次刷新时间。
|
||
# 否则下次 timeout 会变成 0，导致疯狂空转。
|
||
last_flush = time.monotonic()
|
||
# 【核心修复 2】：跳出本次循环，避免重复执行下面的 `if should_flush` 判断
|
||
continue
|
||
|
||
# 满了或超时了就发
|
||
should_flush = (
|
||
len(batch) >= self._batch_size or
|
||
(batch and time.monotonic() - last_flush >= self._flush_interval)
|
||
)
|
||
if should_flush:
|
||
self._send(batch)
|
||
batch = []
|
||
last_flush = time.monotonic()
|
||
|
||
# 退出前 flush 剩余
|
||
if batch:
|
||
self._send(batch)
|
||
|
||
def _send(self, batch):
|
||
"""发送一批日志到 db 服务。失败时静默丢弃。"""
|
||
data = json.dumps({'logs': batch}).encode('utf-8')
|
||
req = urllib.request.Request(
|
||
self._url, data=data,
|
||
headers={'Content-Type': 'application/json'},
|
||
method='POST'
|
||
)
|
||
try:
|
||
with urllib.request.urlopen(req, timeout=_TIMEOUT) as resp:
|
||
resp.read() # 消费响应体
|
||
except (urllib.error.URLError, urllib.error.HTTPError, OSError):
|
||
# fail-safe：日志服务不可用，静默丢弃，不影响业务
|
||
pass
|
||
except Exception:
|
||
pass
|
||
|
||
def _shutdown(self):
|
||
"""进程退出时 flush 剩余日志。"""
|
||
self._stop.set()
|
||
self.flush()
|
||
|
||
|
||
# ── 全局单例 ──────────────────────────────────────────────────────────
|
||
|
||
logs = LogClient()
|