Introduction using the psutil module (https: pypipythonorgpypipsutil), you can conveniently monitor performance parameters such as CPU, memory, disk IO, and network bandwidth of the system, whether the following code is used to monitor the CPU resource consumption of a specific program, print the monitoring data, and finally draw an introduction
Psutil module (https://pypi.python.org/pypi/psutil/) can be very convenient to monitor the system CPU, memory, disk I/O, network bandwidth and other performance parameters, the following code is to monitor a specific program of CPU resource consumption, print the monitoring data, display the drawing, and save it as a backup of the specified PDF document.
Demo code
#!/usr/bin/env python# -*- coding: utf-8 -*-'''Copyright (C) 2015 By Thomas Hu. All rights reserved.@author : Thomas Hu (thomashtq#163.com)@version: 1.0@created: 2015-7-14'''import matplotlib.pyplot as pltimport psutil as psimport osimport timeimport randomimport collectionsimport argparseclass ProcessMonitor(object): def __init__(self, key_name, fields, duration, interval): self.key_name = key_name self.fields = fields self.duration = float(duration) self.inveral = float(interval) self.CPU_COUNT = ps.cpu_count() self.MEM_TOTAL = ps.virtual_memory().total / (1024 * 1024) self.procinfo_dict = collections.defaultdict(dict) def _get_proc_info(self, pid): try: proc = ps.Process(pid) name = proc.name() # If not contains the key word, return None if name.find(self.key_name) == -1: return None pinfo = { "name": name, "pid" : pid, } # If the field is correct, add it to the process information dictionary. for field in self.fields: if hasattr(proc, field): if field == "cpu_percent": pinfo[field] = getattr(proc, field)(interval = 0.1) / self.CPU_COUNT elif field == "memory_percent": pinfo[field] = getattr(proc, field)() * self.MEM_TOTAL / 100 else: pinfo[field] = getattr(proc, field)() if pid not in self.procinfo_dict: self.procinfo_dict[pid] = collections.defaultdict(list) self.procinfo_dict[pid]["name"] = name for field in self.fields: self.procinfo_dict[pid][field].append(pinfo.get(field, 0)) print(pinfo) return pinfo except: pass return None def monitor_processes(self): start = time.time() while time.time() - start < self.duration: try: pids = ps.pids() for pid in pids: self._get_proc_info(pid) except KeyboardInterrupt: print("Killed by user keyboard interrupted!") return def _get_color(self): color = "#" for i in range(3): a = hex(random.randint(0, 255))[2:] if len(a) == 1: a = "0" + a color += a return color.upper() def draw_figure(self, field, pdf): # Draw each pid line for pid in self.procinfo_dict: x = range(len(self.procinfo_dict[pid][field])) #print x, self.procinfo_dict[pid][field] plt.plot(x, self.procinfo_dict[pid][field], label = "pid" + str(pid), color = self._get_color()) plt.xlabel(time.strftime("%Y-%m-%d %H:%M:%S")) plt.ylabel(field.upper()) plt.title(field + " Figure") plt.legend(loc = "upper left") plt.grid(True) plt.savefig(pdf, dpi = 200) plt.show()def Main(): parser = argparse.ArgumentParser(description='Monitor process CPU and Memory.') parser.add_argument("-k", dest='key', type=str, default="producer", help='the key word of the processes to be monitored(default is "producer")') parser.add_argument("-d", dest='duration', type=int, default=60, help='duration of the monitor to run(unit: seconds, default is 60)') parser.add_argument('-i', dest='interval', type=float, default=1.0, help='interval of the sample(unit: seconds, default is 1.0)') args = parser.parse_args() fields = ["cpu_percent", "memory_percent"] #print args.key, args.duration, args.interval pm = ProcessMonitor(args.key, fields, args.duration, args.interval) pm.monitor_processes() pm.draw_figure("cpu_percent", "cpu.pdf") pm.draw_figure("memory_percent", "mem.pdf")if __name__ == "__main__": Main()
Output result demonstration diagram