今天將介紹讀取 /proc/loadavg 的內容,並將資料寫入
rrd file的Python程式.
我們看一下 /proc/loadavg 的內容

0.91 0.64 0.50 1/416 4044

所謂load 就是Job 還在Run Queue裡面尚未執行,或是等待IO,數值越大代表系統越多工作
等待;前三個欄位分別是 1分鐘,5分鐘,15分鐘的負荷平均值,
第四個欄位代表 正在執行的processes與休眠的processes.用 / 分隔.
第五個是剛執行的 process 的 pid.

接著就是 Python 程式碼

#!/usr/bin/env python
# ---------------------------------
# Python RRDTool
# Update LoadAvg and Processes RRD
# ---------------------------------

import rrdtool
import time

def get_loads():
    """ get system avg. load and process status"""
    loads_dict = {}
    loads_list = ('load1', 'load5', 'load15', 'runproc', 'slpproc')
    with open('/proc/loadavg', 'r') as loads:
        sysloads = loads.readline()
        sysloads = sysloads.split()
        load1 = float(sysloads[0])
        load5 = float(sysloads[1])
        load15 = float(sysloads[2])
        procs = sysloads[3]
        runproc , slpproc = procs.split('/')
        runproc = int(runproc)
        slpproc = int(slpproc)
        loads_dict = dict(zip(loads_list, (load1, load5, load15, runproc, slpproc)))
    
    return loads_dict
#
def update_loads_rrd(rrdfile):
    """ update loads rrd file """
    loads_dict = get_loads()
    load1 = loads_dict['load1']
    load5 = loads_dict['load5']
    load15 = loads_dict['load15']
    runproc = loads_dict['runproc']
    slpproc = loads_dict['slpproc']
    totproc = runproc + slpproc
    rrdtool.update(rrdfile, 'N:' + `load1` + ':' + `load5` + ':' + \
                   `load15` + ':' + `totproc` + ':' + `slpproc` + \
                   ':' + `runproc`)
    
        
#
if __name__ == '__main__':
    while 1:
        update_loads_rrd('sysload.rrd')
        time.sleep(60)

放到背景執行,會每分鐘update資料.