我正在运行测试aiohttp webserver：
#!/usr/bin/env python3
from aiohttp import web
import time
import asyncio
import random
import string
import logging
logger = logging.getLogger('webserver')
logger.setLevel(logging.INFO)
handler = logging.StreamHandler()
formatter = logging.Formatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')
handler.setFormatter(formatter)
logger.addHandler(handler)
def randomword(length):
return ''.join(random.choice(string.ascii_lowercase) for i in range(length))
async def hello(request):
sleeptime = 10 # seconds
requestid = randomword(5)
logger.info("Request id: {} received - will sleep for {}".format(requestid, sleeptime))
await asyncio.sleep(sleeptime)
logger.info("Request id: {} wakeup and respond".format(requestid))
return web.json_response()
# create app and register with request handler
app = web.Application()
app.router.add_get('/', hello)
web.run_app(app, port='15000')
预计这将能够同时处理多个传入请求，而无需在处理下一个请求之前等待一个请求完成 . 但是，它没有这样做 - 处理是顺序的（注意第2个/第3个请求的开始时间是在之前的请求完成之后）：
======== Running on http://0.0.0.0:15000/ ========
(Press CTRL+C to quit)
2016-10-23 02:28:23,911 - webserver - INFO - Request id: hpzdo received - will sleep for 10
2016-10-23 02:28:33,922 - webserver - INFO - Request id: hpzdo wakeup and respond
2016-10-23 02:28:33,947 - webserver - INFO - Request id: pdfbs received - will sleep for 10
2016-10-23 02:28:43,957 - webserver - INFO - Request id: pdfbs wakeup and respond
2016-10-23 02:28:43,971 - webserver - INFO - Request id: dmkav received - will sleep for 10
2016-10-23 02:28:53,982 - webserver - INFO - Request id: dmkav wakeup and respond
我希望每个请求的开始时间几乎相同（取决于我提交它们的速度;这里我会在一两秒钟内手动提交） .
1 回答
你的测试技术出了点问题 .
对服务器运行wrk工具会产生不同的结果 .
要运行的命令：
服务器输出：