How to Make a Process Monitor in Python?

A process monitor is a tool that displays the system information like processes, memory, network, and other stuff. There are plenty of tools available, but we can make our own process monitor using Python. In Python, there is a module called psutil that we can use to grab various information about our system

Modules Needed

• psutil: Type the below command in the terminal to install this module.

python3 -m pip install psutil 

• Prettytable: To print the data on console, we can use a formatter module PrettyTable:

python3 -m pip install prettytable

Using psutil

psutil provides lots of features to monitor the system. We will see some of them in brief:

• First, we need to import psutil:

import psutil

• List the process ids:

psutil.pids()  # [1,2,.....4352]

• Fetch process information:

process_id = 1
psutil.Process(process_id)  
# psutil.Process(pid=1, name='systemd', status='sleeping', started='19:49:25')

• We can access various keys of this process:

process = psutil.Process(process_id)
process.name()
process.status()

• Accessing battery status:

psutil.sensors_battery()    
psutil.sensors_battery().percent

• Accessing Network Interfaces:

psutil.net_if_stats()  
psutil.net_if_stats()['wlo1'].isup    # True

• We can also check the memory:

psutil.virtual_memory()
psutil.virtual_memory().total    # 8180498432 (In Bytes)
psutil.virtual_memory().used    # 2155720704
psutil.virtual_memory().available   # 5563060224

Now that we know some basic features, we can implement the process monitor. Create a new python file and add the following code in it. The code below works on Linux distributions. For other operating systems, some functions may slightly differ.

Approach:

• Import the required packages.
• Clear the console using the call() function of the subprocess module. We can use the ‘clear’ or ‘cls’ command depending on OS.
• Fetch the battery information
• Fetch the network information and print it as PrettyTable
• Fetch the memory information
• Fetch the process information
• Create a delay. We have created a 1-second delay using time.sleep(1)
• Press CTRL+C to stop the program.

Below is the implementation:

Python3

# Import the required libraries
import psutil
import time
from subprocess import call
from prettytable import PrettyTable

# Run an infinite loop to constantly monitor the system
while True:

    # Clear the screen using a bash command
    call('clear')

    print("==============================Process Monitor\
    ======================================")

    # Fetch the battery information
    battery = psutil.sensors_battery().percent
    print("----Battery Available: %d " % (battery,) + "%")

    # We have used PrettyTable to print the data on console.
    # t = PrettyTable(<list of headings>)
    # t.add_row(<list of cells in row>)

    # Fetch the Network information
    print("----Networks----")
    table = PrettyTable(['Network', 'Status', 'Speed'])
    for key in psutil.net_if_stats().keys():
        name = key
        up = "Up" if psutil.net_if_stats()[key].isup else "Down"
        speed = psutil.net_if_stats()[key].speed
        table.add_row([name, up, speed])
    print(table)

    # Fetch the memory information
    print("----Memory----")
    memory_table = PrettyTable(["Total(GB)", "Used(GB)",
                                "Available(GB)", "Percentage"])
    vm = psutil.virtual_memory()
    memory_table.add_row([
        f'{vm.total / 1e9:.3f}',
        f'{vm.used / 1e9:.3f}',
        f'{vm.available / 1e9:.3f}',
        vm.percent
    ])
    print(memory_table)

    # Fetch the 10 processes from available processes that has the highest cpu usage
    print("----Processes----")
    process_table = PrettyTable(['PID', 'PNAME', 'STATUS',
                                 'CPU', 'NUM THREADS', 'MEMORY(MB)'])

    proc = []
    # get the pids from last which mostly are user processes
    for pid in psutil.pids()[-200:]:
        try:
            p = psutil.Process(pid)
            # trigger cpu_percent() the first time which leads to return of 0.0
            p.cpu_percent()
            proc.append(p)

        except Exception as e:
            pass

    # sort by cpu_percent
    top = {}
    time.sleep(0.1)
    for p in proc:
        # trigger cpu_percent() the second time for measurement
        top[p] = p.cpu_percent() / psutil.cpu_count()

    top_list = sorted(top.items(), key=lambda x: x[1])
    top10 = top_list[-10:]
    top10.reverse()

    for p, cpu_percent in top10:

        # While fetching the processes, some of the subprocesses may exit
        # Hence we need to put this code in try-except block
        try:
            # oneshot to improve info retrieve efficiency
            with p.oneshot():
                process_table.add_row([
                    str(p.pid),
                    p.name(),
                    p.status(),
                    f'{cpu_percent:.2f}' + "%",
                    p.num_threads(),
                    f'{p.memory_info().rss / 1e6:.3f}'
                ])

        except Exception as e:
            pass
    print(process_table)

    # Create a 1 second delay
    time.sleep(1)

Output: