Thread: Looking for shell script to Monitor CPU usage and send mail once it exceed 75%

Dear Group,

I'm looking for shell script to monitor cpu usage in background and once it exceeded 75% it should send mail.

Thanks in advance.

Hi,

Please check below script

PHP Code:

#!/bin/bash
# Date      = Thu Apr 12 16:52:05 IST 2012
# author = Rahul Patil
# Description = Script to check cpu usage and if above threshold issue a linux based command. 
# This Script only checks "user cpu time (us - userspace (doing userspace stuff)) 
#
# Script log file
LOG="/var/log/cpu.log"

# Define Threshold
THRESHOLD=75  ## in percentage

# wait time(in seconds) after sending mail 
WAIT=60

while :
do
# Obtain the cpu usage
thiscpu_usage=$(top -b -n 1  | awk -F'[:,]' '/^Cpu/{sub("\\..*","",$2); print $2}')

if [ "$thiscpu_usage" -ge "$THRESHOLD" ]; then

# Send Mail To Systems Admins When CPU usage is above Threshold

SUBJECT="WARNING CPU USAGE HIGH"

TO="nobody@emailcompany.com"

# This is a temp file created for the email message....
MESSAGE="/tmp/message.txt"

ps -eo "%cpu,%mem,etime,args" | sort -nr -k 1 | grep -v CPU | head -10 >>  $MESSAGE

echo "Place the body of your message here" >> $MESSAGE
echo  "Time: `date`" >> $MESSAGE

mail -s "$SUBJECT" "$TO" < $MESSAGE

rm $MESSAGE

echo "$(tput bold)CPU USAGE   $(tput setaf 1)CRITICLE@@$(date)@@Usage:$thiscpu_usage $(tput sgr 0)" | tee -a $LOG

sleep $WAIT

else

echo "$(tput bold)CPU usage is OK @@$(date)@@ $(tput setaf 1)Usage:$thiscpu_usage$(tput sgr 0)" | tee -a $LOG
fi
done 


For testing this script use this link or you can simply run below command to generate high cpu usage

HTML Code:

while : ; do : ;done

we can run any script in background using below command

HTML Code:

nohup bash /path_of_script_file &

we can stop script using following command

HTML Code:

fuser -k /path_of_script

also you can check logs

HTML Code:

tail -f /var/log/cpu.log

for configuring mail utility use this link

reference link

hi,

`grep | awk | awk' is really ugly !
what's the use of `tail -1'? I don't understand why output two lines to take the last one.

Code:

top -b -n 1  | awk -F'[:,]' '/^Cpu/{sub("\\..*","",$2); print $2}'

HTH

Hi, Watael

HTML Code:

top -b -n 1  | awk -F'[:,]' '/^Cpu/{sub("\\..*","",$2); print $2}'

thank for giving above trick

Originally Posted by Watael
what's the use of `tail -1'? I don't understand why output two lines to take the last one.

when i use "top -b -n 1" it was giving me same value during testing this script, but now its working fine with your awk trick

Code:

#!/bin/bash

cat <<EOF >~/.toprc
RCfile for "top with windows"    # shameless braggin'
Id:a, Mode_altscr=0, Mode_irixps=1, Delay_time=3.000, Curwin=0
Def   fieldscur=aehioqtwKNMbcdfgjplrsuvyzX
   winflags=42297, sortindx=10, maxtasks=0
   summclr=1, msgsclr=1, headclr=3, taskclr=1
Job   fieldscur=ABcefgjlrstuvyzMKNHIWOPQDX
   winflags=62777, sortindx=0, maxtasks=0
   summclr=6, msgsclr=6, headclr=7, taskclr=6
Mem   fieldscur=ANOPQRSTUVbcdefgjlmyzWHIKX
   winflags=62777, sortindx=13, maxtasks=0
   summclr=5, msgsclr=5, headclr=4, taskclr=5
Usr   fieldscur=ABDECGfhijlopqrstuvyzMKNWX
   winflags=62777, sortindx=4, maxtasks=0
   summclr=3, msgsclr=3, headclr=2, taskclr=3
EOF


trigger=70
tmpfile=$(mktemp)

#having this script in crontab could be a good idea, instead of using a never ending while loop.
#while :; do #uncomment to execute for ever

top -b -n1 | awk -v trg="$trigger" 't >= trg && NR > 4 && NR < 15
   NR == 2 {
      t = $0
      sub("[^ ]* *","",t)
      sub("\..*","",t)
      if(t >= trg){print}
   }' > $tmpfile

[ -s "$tmpfile" ] && mail -s "overloadCPU" root@localhost < $tmpfile
[ -f "$tmpfile" ] && rm "$tmpfile"

#sleep 60 #idem
#done #idem

rm ~/.toprc

I didn't write the .toprc file myself, I only recorded from `top', and copy/paste it in the script.

Hi Rahul,

First thanks for your replay.
in your script what I understood after ran the script that I should wait till the THRESHOLD reached 75 but what I'm looking for is once the CPU% reached 75% email should be trigger, and I would like to know how to make this script run in the cron job continually.

Hi clfever,
you can run this script using cron but for that you need to comment while loop.
for cron, add following entry in "/etc/crontab" then once restart cron service
#Ex.
# run every 5 minut
*/5 * * * * root /absolute_path_of_script_file

for more information about cron use this link

try using

memUsage=`sar -q 1 | tail -1 | awk '{ print "" $3}' | sed 's/%//g'`

this doesnt give the memory usage in percentage terms but if it could be of use.

Thanks & Regards,
vivek

Originally Posted by clfever

Dear Group,

I'm looking for shell script to monitor cpu usage in background and once it exceeded 75% it should send mail.

Thanks in advance.

Originally Posted by Rahul.Patil

Hi,
Please check below script

Thanks for your script, based onto your script I wrote own tuned for muliple CPUs and tuned 'top' command for OS FreeBSD.
by this link at my profile present my publication with description.

Code:

#!/usr/local/bin/bash
#monitoring idle state per processor core
# Script log file
LOG="/var/log/cpu.log"


# Define Threshold
THRESHOLD=5  ## in percentage idle


TO="postmaster@example.com"
MESSAGE="/tmp/monitor-cpu-message.txt"


# wait time(in seconds) after sending mail
WAITMAIL=300
# wait time(in seconds) after loop
WAIT=8
# wait time (in seconds) in critical zone
INITWINDOW=120


#fill init value for all CPUs
for index in 0 1 2 3
do
 WINDOW[index]=$INITWINDOW
done


function checkwindow() {
    if [ $1 -eq 0 ]; then
        IDLE=$3
        let "usage=100-IDLE"
        SUBJECT="WARNING CPU$2 USAGE HIGH"
        ps -o "%cpu,args" | sort -nr -k 1  | head -10 >>  $MESSAGE
        echo "." >> $MESSAGE
        echo "Usage of CPU$2 is $usage% during $INITWINDOW sec." >> $MESSAGE
        echo  "Time: `date`" >> $MESSAGE
        mail -s "$SUBJECT" "$TO" < $MESSAGE
        rm $MESSAGE
        echo "CPU$2 USAGE CRITICLE @@$(date)@@ Usage:$usage" | tee -a $LOG
        sleep $WAITMAIL
    fi
}


while :
do
CPU=0
  for thiscpu_usage in $(top -b -d 2 -I -P  | awk '/^CPU/{sub(/\.+% idle$/,"");  print int($11)}') ; do
    if [ "$thiscpu_usage" -le "$THRESHOLD" ]; then
        let "WINDOW[CPU]=WINDOW[CPU]-2-WAIT"
    #    echo "CPU$CPU $thiscpu_usage ${WINDOW[$CPU]}"
        checkwindow ${WINDOW[$CPU]} $CPU $thiscpu_usage
    else
        WINDOW[$CPU]=$INITWINDOW
    #    echo "CPU$CPU usage is OK @@$(date)@@ Idle:$thiscpu_usage%" | tee -a $LOG
    fi
  let "CPU=CPU+1"
  done
sleep $WAIT
done