[LINUX] [Note] A shell script that checks the CPU usage of a specific process in a while loop.

0. Introduction

Share the shell script created to periodically monitor the process with the top command. Specifically, it is as simple as continuing to output the execution result of the top command to the log file in a while loop.

1. Verification environment, etc.

--Local pc

• ubuntu20.04

2. Execution result of top command

First, share the execution result of the top command to be a shell script as a sample.

$ top -n 1
top - 00:58:08 up 55 min,  1 user,  load average: 0.67, 0.58, 0.64
Tasks: 329 total,   1 running, 328 sleeping,   0 stopped,   0 zombie
%Cpu(s):  4.6 us,  0.8 sy,  0.0 ni, 93.1 id,  0.0 wa,  0.0 hi,  1.5 si,  0.0 st
MiB Mem :  15662.8 total,   9716.0 free,   2467.8 used,   3478.9 buff/cache
MiB Swap:   2048.0 total,   2048.0 free,      0.0 used.  11582.7 avail Mem 

    PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND                                                                                                                             
   9095 gkz       20   0 4616108 107196  70536 S  25.0   0.7   0:00.37 chrome                                                                                                                              
   2487 gkz       20   0 4730532 335940 156620 S   6.2   2.1   2:42.61 gnome-shell                                                                                                                         
   4102 gkz       20   0  321656  12708  11268 S   6.2   0.1   0:05.56 ibus-engine-moz                                                                                                                     
   7243 gkz       20   0  961388  52248  37448 S   6.2   0.3   0:05.53 gnome-terminal-                                                                                                                     
      1 root      20   0  168164  11936   8272 S   0.0   0.1   0:10.67 systemd    

The top command can limit the number of screen (frame?) Updates by using the n option. In this article, in order to periodically monitor the CPU usage of any process, we will pipe the result of the top command and use grep to narrow down the output result of a specific process.

3. Source code

toploop.sh

#!/bin/bash

# debug mode
# bash -x toploop.sh $process

# your monitored process, e.g. dockerd
process=$1

starttime=`date +"%Y%m%d_%H%M%S"`
filename=`echo "log/top.log.$process.$starttime"`

while true
do
  sleep 5
  output=`top -n 1| grep $process`
  val=`echo $output | awk '{print $13}'`
  if [ "$val" == "$process" ]; then
    num=`echo "$output" | grep -oP "[S|R]\s+[0-9]*[.]?[0-9]" | awk '{print $2}'`
    date +"%Y%m%d_%H%M%S" >> $filename
    echo "$process:  $num%" >> $filename
  else
    date +"%Y%m%d_%H%M%S" >> $filename
    echo "$process: notfound" >> $filename
  fi
done

4. Execution method

4-1. Execute the command in the background with & at the end of the command

$  . toploop.sh chrome&
[1] 11175

4-2. After executing the shell script, follow the log file update with the tail command.

• In the shell script, the format of the log file name is top.log. $ {Process}. $ {Starttime}.

$ tail -f log/top.log.chrome.20200526_011531
20200526_011602
chrome: notfound
20200526_011608
chrome:  6.2%

• After monitoring the process, don't forget to execute the shell script executed in 4-1 with kill $ {bash PID}.

5. Reference

• top(1) - Linux manual page

P.S. I also do Twitter, so if you follow me, I'll be happy to cry :)

@gkzvoice