Question

我的python代码调用了REST上的localhost服务器。服务器本身会记录指标（即处理时间），并在响应中发回自己的测量结果以进行调试。

由于某种原因，我不知道从python代码调用localhost服务器会有很大的开销。我已经用一些简单的代码测试了时间：

import logging
import requests
import time
import os
os.environ['no_proxy'] = '127.0.0.1,localhost'
# asserting that the connection is being re-used - this prints it out
logging.basicConfig(level=logging.DEBUG)
# re-use the session
s = requests.Session()
# try variations of host to see if there is a difference
localhost = "localhost"
local_ip = "127.0.0.1"

# https://github.com/requests/requests/issues/879
NO_PROXY = {
    'no': 'pass',
}

def measure():
    url = "http://{}:8888/endpoint".format(localhost)
    start = time.time()
    response = s.get(url, proxies=NO_PROXY)
    end = time.time()
    print("took {} ms".format((end-start)*1000))

    url = "http://{}:8888/endpoint".format(local_ip)
    start = time.time()
    response = s.get(url, proxies=NO_PROXY)
    end = time.time()
    print("took {} ms".format((end-start)*1000))

for i in range(0,1000):
    make_request()
    ping()

日志看起来像

DEBUG:urllib3.connectionpool:Starting new HTTP connection (1): localhost:8888
DEBUG:urllib3.connectionpool:http://localhost:8888 "GET /endpoint HTTP/1.1" 200 4
took 7.76600837708 ms
DEBUG:urllib3.connectionpool:Starting new HTTP connection (1): 127.0.0.1:8888
DEBUG:urllib3.connectionpool:http://127.0.0.1:8888 "GET /endpoint HTTP/1.1" 200 4
took 4.73213195801 ms
DEBUG:urllib3.connectionpool:http://localhost:8888 "GET /endpoint HTTP/1.1" 200 4
took 1.47581100464 ms
DEBUG:urllib3.connectionpool:http://127.0.0.1:8888 "GET /endpoint HTTP/1.1" 200 4
took 1.25908851624 ms
DEBUG:urllib3.connectionpool:http://localhost:8888 "GET /endpoint HTTP/1.1" 200 4
took 1.25503540039 ms
DEBUG:urllib3.connectionpool:http://127.0.0.1:8888 "GET /endpoint HTTP/1.1" 200 4
took 1.21402740479 ms
DEBUG:urllib3.connectionpool:http://localhost:8888 "GET /endpoint HTTP/1.1" 200 4
took 1.31893157959 ms
DEBUG:urllib3.connectionpool:http://127.0.0.1:8888 "GET /endpoint HTTP/1.1" 200 4
took 1.54900550842 ms

我还打印了标题，以了解此请求中正在传输多少数据（总共约300个字节）：

{'Content-Length': '138', 'Accept-Encoding': 'gzip, deflate', 'Accept': '*/*', 'User-Agent': 'python-requests/2.19.1','Connection': 'keep-alive', 'debug': 'true', 'Content-Type': 'application/json'}

我看到初始调用很慢，但是后来的调用正在重新使用连接（明显更快），但是仍然，当我与服务器日志关联时，我发现这1.5毫秒是一个相对较长的时间。服务器的速度显着提高，花费了不到0.1毫秒的时间，这使我相信本地主机调用要付出一些严厉的代价。

此外，我们看到"content-length"很小-在这种情况下为138

Just for comparison with regular "ping" command:

ping localhost
PING localhost (127.0.0.1): 56 data bytes
64 bytes from 127.0.0.1: icmp_seq=0 ttl=64 time=0.051 ms
64 bytes from 127.0.0.1: icmp_seq=1 ttl=64 time=0.096 ms
64 bytes from 127.0.0.1: icmp_seq=2 ttl=64 time=0.073 ms
64 bytes from 127.0.0.1: icmp_seq=3 ttl=64 time=0.089 ms
64 bytes from 127.0.0.1: icmp_seq=4 ttl=64 time=0.120 ms
64 bytes from 127.0.0.1: icmp_seq=5 ttl=64 time=0.078 ms
64 bytes from 127.0.0.1: icmp_seq=6 ttl=64 time=0.097 ms

所以快20倍。我意识到它使用的是完全不同的协议，但仍仅作为参考。

用python调用本地主机服务器会产生大量开销

0 个答案: