C++版本的MsgClient

This commit is contained in:
qing committed 2026-07-26 21:48:14 +08:00
1 parent 3d2278f6b7
commit 6efb792133
4 files changed
+545 -156

No files matched your search

+185 -152
View File
@@ -1,152 +1,185 @@
"""
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:
pass # 超时,没有新日志
# 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()
"""
统一日志客户端 SDK — 各服务复制即用。
设计要点:
- 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,绝不影响业务
- 缓冲批量:队列满 50 条或超时 2 秒时自动 flush
复制到目标服务目录后,按需修改 service_name 即可::
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 os as _os
import sys as _sys
import threading
import queue
import time
import atexit
from datetime import datetime, timezone
_sys.path.insert(0, _os.path.join(_os.path.dirname(__file__), ".."))
from pymsg import MsgClient # noqa: E402
# ── 配置 ──────────────────────────────────────────────────────────────
_DBV2_ADDR = '127.0.0.1'
_DBV2_PORT = 8005
_BATCH_SIZE = 50 # 队列满 50 条时 flush
_FLUSH_INTERVAL = 2.0 # 超时 2 秒强制 flush
_TIMEOUT = 5 # socket 超时(秒)
class LogClient:
"""异步批量日志推送客户端。
通过 MsgClient(pymsg TCP 协议)逐条 INSERT 到 dbv2 的 Logs 表。
service_name 用于在 dbv2 连接中标识来源,各服务复制时自行修改。
"""
def __init__(self, service_name="example",
batch_size=_BATCH_SIZE,
flush_interval=_FLUSH_INTERVAL):
self._service_name = service_name
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:
# 超时无新日志:如果 batch 中已有积攒的条目,强制 flush
if batch:
self._send(batch)
batch = []
last_flush = time.monotonic()
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):
"""通过 pymsg 使用 execute 路由逐条 INSERT 到 dbv2 Logs 表。"""
now = datetime.now(timezone.utc).isoformat()
try:
client = MsgClient(_DBV2_ADDR, _DBV2_PORT, self._service_name + "-log")
for entry in batch:
extra_val = entry.get("extra")
payload = {
"_route": "execute",
"action": "insert",
"table": "Logs",
"values": {
"ts": now,
"service": entry.get("service"),
"level": entry.get("level"),
"message": entry.get("message"),
"trace_id": entry.get("trace_id"),
"extra": json.dumps(extra_val, ensure_ascii=False)
if extra_val is not None else None,
},
}
resp = client.call(json.dumps(payload), timeout=_TIMEOUT)
resp_obj = json.loads(resp)
if not resp_obj.get("success"):
import sys
print(f"[log_client] INSERT 失败: {resp_obj.get('error')}",
file=sys.stderr)
except Exception as e:
import sys
print(f"[log_client] 发送失败: {e}", file=sys.stderr)
def _shutdown(self):
"""进程退出时 flush 剩余日志。"""
self._stop.set()
self.flush()
# ── 全局单例 ──────────────────────────────────────────────────────────
# 各服务复制时,把 "example" 改为自己的服务名即可。
# 例如 dnfs/log_client.py → logs = LogClient(service_name="dnfs")
# 例如 smtp/log_client.py → logs = LogClient(service_name="smtp")
logs = LogClient()