用Python写了个简单的监控进程的脚本monitor190620.py，记录进程的CPU利用率和内存利用率到文件pid.csv中，分析进程运行数据用图表展示。
脚本的工作原理是这样的：脚本读取配置文件config.ini，读取预先配置好的进程的PID，之所以要用配置文件，是为了方便给需要的朋友，你只需要修改PID就可以了，源代码就不需要修改了。具体的看代码中的注释吧。下面的是配置文件 config.ini
[MonitorProcessID] ProcessID: 7544
上面可以根据你的需求配置不同的进程ID，我是需要监控 JMeter这个程序，通过 jps -l 查看该进程的PID，填写配置文件内容。
下面来看看monitor190620.py的代码（该代码在CentOS下运行时需要改正一行代码：from ConfigParser import ConfigParser）：
# -*- encoding: utf-8 -*-
import psutil
import csv
import time
from configparser import ConfigParser
CONFIGFILE = 'config.ini'
config = ConfigParser()
config.read(CONFIGFILE)
ProcessID = config.get('MonitorProcessID', 'ProcessID')
# 读取配置文件中的进程ID，这个参数可以在配置文件中修改
proc = psutil.Process(int(ProcessID))
#print(ProcessID,"% .2f% %"%(proc.memory_percent()),"% .2f% %"%(proc.cpu_percent()),proc.name(),proc.exe())
f = open('pid.csv', 'w')
f.truncate()
f_csv = csv.writer(f)
flag = 1
while flag < 10:
row = []
row = [ProcessID,'%.2f'%(proc.memory_percent()),'%.2f'%(proc.cpu_percent()),time.strftime("%H:%M:%S", time.localtime()),proc.name(),proc.exe()]
f_csv.writerow(row)
flag += 1
time.sleep(4)
f.close()
执行后，生成的 pid.csv 的文件内容为：
7544,3.12,0.00,15:39:12,java.exe,C:\Program Files (x86)\Common Files\Oracle\Java\javapath_target_119167744\java.exe 7544,3.12,0.00,15:39:16,java.exe,C:\Program Files (x86)\Common Files\Oracle\Java\javapath_target_119167744\java.exe 7544,3.12,0.00,15:39:20,java.exe,C:\Program Files (x86)\Common Files\Oracle\Java\javapath_target_119167744\java.exe 7544,3.12,0.00,15:39:24,java.exe,C:\Program Files (x86)\Common Files\Oracle\Java\javapath_target_119167744\java.exe 7544,3.12,0.00,15:39:28,java.exe,C:\Program Files (x86)\Common Files\Oracle\Java\javapath_target_119167744\java.exe 7544,3.12,0.00,15:39:32,java.exe,C:\Program Files (x86)\Common Files\Oracle\Java\javapath_target_119167744\java.exe 7544,3.12,0.00,15:39:36,java.exe,C:\Program Files (x86)\Common Files\Oracle\Java\javapath_target_119167744\java.exe 7544,3.12,0.00,15:39:40,java.exe,C:\Program Files (x86)\Common Files\Oracle\Java\javapath_target_119167744\java.exe 7544,3.12,0.00,15:39:44,java.exe,C:\Program Files (x86)\Common Files\Oracle\Java\javapath_target_119167744\java.exe
编写 show190620.py ，读取 pid.csv 的文件内容，做统计分析，并将分析结果可视化为图表：
import csv
import pandas as pd
import matplotlib.pyplot as plt
time = []
cpu = []
mem = []
with open('pid.csv')as f:
f_csv = csv.reader(f)
for row in f_csv:
if len(row): #判断是不是空行，csv中隔一行记录运行的信息
time.append(row[3])
cpu.append(float(row[2]))
mem.append(float(row[1]))
else:
continue
print(time)
print(cpu)
print(mem)
s_cpu = pd.Series(cpu)
s_mem = pd.Series(mem)
cpu_mean = s_cpu.mean()
mem_mean = s_mem.mean()
print('cpu利用率平均值是：%f' % cpu_mean)
print('内存利用率平均值是：%f' % mem_mean)
plt.plot(time,cpu, c='r')
plt.plot(time,mem, c='b')
plt.show()
返回信息：
['15:39:12', '15:39:16', '15:39:20', '15:39:24', '15:39:28', '15:39:32', '15:39:36', '15:39:40', '15:39:44'] [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0] [3.12, 3.12, 3.12, 3.12, 3.12, 3.12, 3.12, 3.12, 3.12] cpu利用率平均值是：0.000000 内存利用率平均值是：3.120000
并显示图表：
参考：
https://www.cnblogs.com/yueminghai/p/6632871.html
https://www.cnblogs.com/xiaobeibei26/p/6481707.html
https://www.cnblogs.com/ratels/p/11023109.html
https://www.cnblogs.com/sen-c7/p/9473224.html
https://blog.****.net/cz505632696/article/details/80025816
https://blog.****.net/katyusha1/article/details/81606175
https://www.runoob.com/python/python-lists.html
https://www.cnblogs.com/AiyaFocus/p/AiyaFocus.html
https://www.cnblogs.com/haiyan123/p/9104143.html