使用locust进行Websocket压力测试和接口压力测试
安装
安装 Locust:
pip3 install locust查看版本
locust -V
# locust 2.5.1基本用法
示例代码
from locust import User, events, task, between
import random
def success_call(name, recvText, total_time):
events.request_success.fire(
request_type="[Success]",
name=name,
response_time=total_time,
response_length=len(recvText)
)
def fail_call(name, total_time):
events.request_failure.fire(
request_type="[Fail]",
name=name,
response_time=total_time,
response_length=0,
exception=Exception(),
)
class MyUser(User):
wait_time = between(1, 5)
@task(1)
def job1(self):
print('This is job 1')
self.sendMsg("job1", 'This is job 2')
@task(2)
def job2(self):
print('This is job 2')
self.sendMsg("job2", 'This is job 2')
def sendMsg(self, methodname, msg):
num = random.randint(0, 2)
time = random.randint(10, 50)
if (num == 0):
fail_call(methodname, time)
else:
success_call(methodname, msg, time);执行
locust -f test4.py --autostart --autoquit 0 -u 1 -r 3 --run-time 10s其中
request_type应该是请求的方式，这里为了方便理解，直接使用的[Success]和[Fail],来区分成功的请求和失败的请求，只有成功或失败的回调被调用后，locust才会对其统计。@task(2)里面的数字是权重。wait_time是两个任务之间的间隔时间，如果不设置则会尽可能快的执行。wait_time = between(1, 5)间隔是1-5秒随机值wait_time = constant(1)间隔固定1秒
注意：
所有的task不能是异步方法。
结果
image-20220116154519702
TaskSet
如果你正在测试一个网站的性能，这个网站是以分层的方式构建的，包括部分和子部分，那么以同样的方式构建负载测试可能会很有用。为此，locust提供了任务集类。它是将执行的任务的集合，与直接在用户类上声明的任务非常相似。
from locust import User, TaskSet, constant,task
class ForumSection(TaskSet):
wait_time = constant(1)
@task(10)
def job1(self):
print("job1----")
@task
def job2(self):
print("job2+++++")
@task
def stop(self):
self.interrupt()
class LoggedInUser(User):
wait_time = constant(5)
tasks = {ForumSection: 2}
@task
def my_task(self):
print("my_task=======")执行
locust -f test5.py --autostart --autoquit 0 -u 1 -r 3 --run-time 10s也可以这样写
class MyUser(User):
@task
class MyTaskSet(TaskSet):
...SequentialTaskSet
SequentialTaskSet是一个任务集，其任务将按照声明的顺序执行。可以在任务集中嵌套顺序任务集，反之亦然。
from locust import User, SequentialTaskSet, constant,task
class MySet(SequentialTaskSet):
wait_time = constant(1)
@task(10)
def job1(self):
print("job1----")
@task
def job2(self):
print("job2+++++")
@task
def stop(self):
print("stop")
self.interrupt()
class LoggedInUser(User):
wait_time = constant(5)
tasks = {MySet: 2}
@task
def my_task(self):
print("my_task=======")执行
locust -f test6.py --autostart --autoquit 0 -u 1 -r 3 --run-time 60s先后顺序
from locust import User, SequentialTaskSet, constant, task
class ASet(SequentialTaskSet):
wait_time = constant(1)
@task(2)
def job1(self):
print("a-1")
@task
def job2(self):
print("a-2")
@task
def stop(self):
print("a-stop")
self.interrupt()
class BSet(SequentialTaskSet):
wait_time = constant(1)
@task(2)
def job1(self):
print("b-1")
@task
def job2(self):
print("b-2")
@task
def stop(self):
print("b-stop")
self.interrupt()
class LoggedInUser(User):
wait_time = constant(1)
tasks = {ASet: 2, BSet: 1}
@task
def my_task(self):
print("c-1")从结果看
a,b,c三者并没有先后的执行循序 ASet和BSet内是按照循序执行的 ASet和BSet执行的概率为2:1
接口压测
在当前目录下创建locustfile.py文件
import random
from locust import HttpUser, task, between
class QuickstartUser(HttpUser):
wait_time = between(1, 3)
host = "http://www.psvmc.cn"
@task
def index_page(self):
self.client.get("/index.html") # 这里的地址需要排除 host 部分
self.client.get("/userlist.json")
@task(3)
def view_item(self):
item_id = random.randint(1, 10000)
self.client.get(f"/search.json?id={item_id}", name="/search.json")
def on_start(self):
self.client.post("/login.json", {"username":"foo", "password":"bar"})运行
locust注意
Locust 会默认查找当前目录下名为
locustfile.py，如果该文件不在当前文件夹或者你取了其他的名字，就需要加上下面的-f参数了。
比如
locust -f locust_files/my_locust_file.py网页上访问
连接WS
压测WS前我们要先看看怎么连接WS
连接WS的库有的是支持异步IO的，项目中我们推荐这样的库，但是压测时还是要选择同步的库。
异步
安装依赖
pip install websockets代码示例
import asyncio
import websockets
import json
import random
async def mytest():
async with websockets.connect('wss://sockettest.xhkjedu.com/ws') as websocket:
num = random.randint(0, 10000000)
msg = {
"b": {"num": num},
"c": 123456,
}
msgstr = json.dumps(msg)
await websocket.send(msgstr)
print(f"↑: {msgstr}")
greeting = await websocket.recv()
print(f"↓: {greeting}")
asyncio.get_event_loop().run_until_complete(mytest())同步
官网地址
https://pypi.org/project/websocket-client/
安装
pip install websocket-client示例
from websocket import create_connection
import json
import random
ws = create_connection("wss://sockettest.xhkjedu.com/ws")
num = random.randint(0, 10000000)
msg = {
"b": {"num": num},
"c": 123456,
}
msgstr = json.dumps(msg)
print("Sending " + msgstr)
ws.send(msgstr)
result = ws.recv()
print("Received '%s'" % result)
ws.close()或者
import websocket
def on_message(ws, message):
print(ws)
print(message)
def on_error(ws, error):
print(ws)
print(error)
def on_close(ws):
print(ws)
print("### closed ###")
websocket.enableTrace(True)
ws = websocket.WebSocketApp("ws://127.0.0.1:8888/track",
on_message=on_message,
on_error=on_error,
on_close=on_close)
ws.run_forever()Websocket压测
Jmeter要测试websocket接口，需要先下载安装一个websocket samplers by peter doornbosch的插件
而locust因为是代码实现，所以可以进行任何的测试，引用相应的库即可。
from locust import User, task, events
import time
from websocket import create_connection
import json
import random
def success_call(name, recvText, total_time):
events.request_success.fire(
request_type="[Success]",
name=name,
response_time=total_time,
response_length=len(recvText)
)
def fail_call(name, total_time, e):
events.request_failure.fire(
request_type="[Fail]",
name=name,
response_time=total_time,
response_length=0,
exception=e,
)
class WebSocketClient(object):
def __init__(self, host):
self.host = host
self.ws = None
def connect(self, burl):
self.ws = create_connection(burl)
def recv(self):
return self.ws.recv()
def send(self, msg):
self.ws.send(msg)
class WebsocketUser(User):
abstract = True
def __init__(self, *args, **kwargs):
super(WebsocketUser, self).__init__(*args, **kwargs)
self.client = WebSocketClient(self.host)
self.client._locust_environment = self.environment
class ApiUser(WebsocketUser):
host = "wss://sockettest.xhkjedu.com/"
@task(1)
def pft(self):
# wss 地址
self.url = 'wss://sockettest.xhkjedu.com/ws'
print("连接前")
start_time = time.time()
try:
self.client.connect(self.url)
print("连接后")
# 发送的订阅请求
num = random.randint(0, 10000000)
msg = {
"b": {"num": num},
"c": 123456,
}
msgstr = json.dumps(msg)
self.client.send(msgstr)
print(f"↑: {msgstr}")
greeting = self.client.recv()
print(f"↓: {greeting}")
except Exception as e:
total_time = int((time.time() - start_time) * 1000)
fail_call("Send", total_time, e)
else:
total_time = int((time.time() - start_time) * 1000)
success_call("Send", "success", total_time)测试
locust -f main.py -u 5000 -r 300网页上访问
或者不用图形化界面
locust -f main.py --autostart --autoquit 0 -u 1 -r 3 --run-time 10s其中
--autostart自动开始 不使用WebUI--autoquit 0和autostart搭配使用，测试完成后多长时间退出，后面的数字单位是秒，如果不设置只能CTRL+C才能退出-u 1最大用户数-r 3每秒创建的用户数，创建用户数和最大用户数一样后就不再创建--run-time 10s压测的执行时间
本文参与 腾讯云自媒体分享计划，分享自作者个人站点/博客。
原始发表：2022-01-16，如有侵权请联系 cloudcommunity@tencent.com 删除
评论
登录后参与评论
推荐阅读
目录