【性能测试】【locust】场景性能测试步骤

场景设计

1. 实现登陆基本功能，输出相应结果，脚本通
2. 多用户实现随机登陆
3. 添加初始化方法on_start: 每个用户只运行一次
4. 添加检查点： catch_responses = True

脚本设计

# 导入对应的库
from locust import HttpUser, task, between
import os
# 任务类
class TestLogin(TaskSet):
    # 定义一个登陆任务
    @task
    def to_login(self):
        """
           1. 登陆需要账号密码，body以json格式
           2. 登陆是post请求 '/v1/token'
           3. 打印res值
        """
        data = {"account": "wangxiao@qq.com", "password": "123456"}
        res = self.client.post("/v1/token", json=data)
        print(res.text)
class WebUser(HttpUser):
    tasks = [TestLogin]
    # 任务执行之间等待的时间，最小2秒，最大5秒
    wait_time = between(2, 5)
     # 被测地址：这个接口是自己写的一个登陆接口，实际可以需要自己抓包，更换真实环境中的地址
    host = "http://127.0.0.1:5000"
if __name__ == '__main__':
    os.system("locust -f testsm_demo.py")
# 设置用户1，每秒启动用户1，确定脚本是通，从打印出来的内容我们可以看出，脚本是通过的（这里你也可以打开locust webUI界面）
[2020-08-19 10:53:19,249] md2bkpyc/INFO/locust.runners: All users hatched: WebUser: 1 (0 already running)
{
  "error_code": 0,
  "token": "eyJhbGciOiJIUzUxMiIsImlhdCI6MTU5NzgwNTYwNCwiZXhwIjoxNTk3ODEyODA0fQ.eyJ1aWQiOjEsInNjb3BlIjpudWxsfQ.GS-lHngBcboPvnipWOl1ad5bkSEYweLwzSMa8fZpYVoz5ZDndV0gJOPvUiu8oiXIZi6AR7amLUcG7KKcBsf0Cg"
}

多用户实现随机登陆

• 定义一组登陆账号
• 导入choice方法，从data中随机选一个登陆

__author__ = 'wangxiao'
# 导入对应的库
from locust import HttpUser, task, between, TaskSet
import os
from random import choice
# 任务类
class TestLogin(TaskSet):
    @task
    def to_login(self):
        data = [{"account": "wangxiao@qq.com", "password": "123456"},
                {"account": "tester06@qq.com", "password": "123456"},
                {"account": "admin@admin.com", "password": "123456"}
                ]
        res = self.client.post("/v1/token", json=choice(data))
        print(res.text)
class WebUser(HttpUser):
    tasks = [TestLogin]
    wait_time = between(2, 5)
    host = "http://127.0.0.1:5000"
if __name__ == '__main__':
    os.system("locust -f testsm_demo.py")

• 设计场景，总共5个用户，每秒启动5个，图表

添加初始化方法on_start

• on_start 与on_stop方法对应，每个虚拟用户执行操作时前会运行一次on_start 推迟执行 on_stop
• 重构方法，把data放到on_start中

__author__ = 'wangxiao'
# 导入对应的库
from locust import HttpUser, task, between, TaskSet
import os
from random import choice
# 任务类
class TestLogin(TaskSet):
    def on_start(self):
        self.data = [{"account": "wangxiao@qq.com", "password": "123456"},
                     {"account": "tester06@qq.com", "password": "123456"},
                     {"account": "admin@admin.com", "password": "123456"}
                     ]
    def on_stop(self):
        print("task结束")
    @task
    def to_login(self):
        res = self.client.post("/v1/token", json=choice(self.data))
        print(res.text)
class WebUser(HttpUser):
    tasks = [TestLogin]
    wait_time = between(2, 5)
    host = "http://127.0.0.1:5000"
if __name__ == '__main__':
    os.system("locust -f testsm_demo.py")

设置检查点： catch_responses = True

• catch_response: 布尔值，True or False, 用于发送请求，作为with语句返回上下文管理器的参数，将根据不同的响应将请求内容标记成功or失败，所以这边用 catch_response来捕捉，注意要用with语句来返回上下文
• success() failure(): 响应类中充当了上下文管理器，他提供了两个方法success() or failure()

• 脚本实例

__author__ = 'wangxiao'
# 导入对应的库
from locust import HttpUser, task, between, TaskSet
import os
from random import choice
# 任务类
class TestLogin(TaskSet):
    def on_start(self):
        self.data = [{"account": "wangxiao@qq.com", "password": "123456"},
                     {"account": "tester06@qq.com", "password": "123456"},
                     {"account": "admin@admin.com", "password": "123456"},
                     {"account": "admin@admin.com", "password": "12334456"}
                     ]
    @task
    def to_login(self):
        with self.client.post("/v1/token", json=choice(self.data), catch_response=True) as response:
            res = response.json()
            if res['error_code'] == 0:
                response.success()
            else:
                response.failure(res['msg'])
    def on_stop(self):
        print("task结束")
class WebUser(HttpUser):
    tasks = [TestLogin]
    wait_time = between(2, 5)
    host = "http://127.0.0.1:5000"
if __name__ == '__main__':
    os.system("locust -f testsm_demo.py")

• 图表分析

• 错误信息提示