Email report about server resource usage based on sar command (crontab ready)
With this crontab-ready script you can check CPU/RAM/IO/Load/HDD usage with monthly averages (except for hard disk usage). Script is based on sar command – and sar is part of sysstat package – so if you dont have it installed – install it (on CentOS: yum install sysstat).
This script is some kind of quick resolution – in case you don’t use Nagios or similar monitoring solutions. Sysstat collect usage snapshots on every 10 minutes by default – but you can increase or decrease this value by changing crontab file: /etc/cron.d/sysstat.
Sar by default (on CentOS) collect information’s in last 30 days (in Ubuntu Server – 14 days) but you an also change this value in sysstat config file /etc/sysconfig/sysstat (line begin with HISTORY)
#!/bin/bash ######################################################################################### # Description: Crontab-ready script for resource usage which will be sent to email # Author: Kompjuteras.com, but you are free to change and use this script however you want # # Crontab example (for running at first day in the month): # 00 01 1 * * /root/scripts/ResourceUsageReport.sh > /tmp/CronResourceUsage.log 2>&1 # ######################################################################################### shopt -s expand_aliases # In case of mail use some smtp variables source /root/.bash_profile # for crontab usage (if you run as a root) ###################################### VARIJABLE ######################################### # Server name in user-friendly format: SomeYourPrefix_HOSTNAME_IPADDRESS # You can hadcode this value if you want (it will be used for email subject or body, like: # SERVER_NAME=KOMPJUTERAS_PROD_SERVER SERVER_NAME="TESTSERVER_$( echo `hostname -s` | tr [:lower:] [:upper:] ) - `ip addr | grep 'state UP' -A2 | grep 'inet ' | tail -n1 | awk '{print $2}' | cut -f1 -d'/'`" # Who need to get this report on email? MAIL_TO="moj@mejl.kom" # Where will be temp logs - best option is /tmp folder LOGFILE_PREFIX=/tmp/SAR # Where are sar achive files? On CentOS is /var/log/sa/ on Ubuntu is /var/log/sysstat/ SAR_LOG_FILE_LOCATION="/var/log/sa" # You you want report per day (very detailed report - day-minute based) # Y for yes and N for N DAYLY_BASED_REPORT=Y # Do you want average values about resource usage check? # WARNING: If you do restart of server - you will get 2 line for that day, average until restart # and average at the end of day # Y for yes and N for N AVERAGES_USAGE_REPORT=Y ########################################################################################## # Check: is sysstat installed. If not - then exit if [ `hash sar 2>/dev/null ; echo $?` -gt 0 ] ; then echo "Instalirajte sysstat program prvo, izlazim!" ; exit 1 ; fi if [ `hash mail 2>/dev/null ; echo $?` -gt 0 ] ; then echo "Instalirajte mailx program prvo, izlazim!" ; exit 1 ; fi # Remove old logs rm -f ${LOGFILE_PREFIX}_*.log # Full log generating - if DAYLY_BASED_REPORT=Y you will get it in separate mail for i in {01..30} do if [ -f ${SAR_LOG_FILE_LOCATION}/sa${i} ] ; then echo "`sar -u -f ${SAR_LOG_FILE_LOCATION}/sa${i} ` | DayOfTheMonth ${i}" | grep -v `hostname` >> ${LOGFILE_PREFIX}_CPU.log echo "`sar -r -f ${SAR_LOG_FILE_LOCATION}/sa${i} ` | DayOfTheMonth ${i}" | grep -v `hostname` >> ${LOGFILE_PREFIX}_RAM.log echo "`sar -S -f ${SAR_LOG_FILE_LOCATION}/sa${i} ` | DayOfTheMonth ${i}" | grep -v `hostname` >> ${LOGFILE_PREFIX}_SWP.log echo "`sar -q -f ${SAR_LOG_FILE_LOCATION}/sa${i} ` | DayOfTheMonth ${i}" | grep -v `hostname` >> ${LOGFILE_PREFIX}_LOAD.log echo "`sar -b -f ${SAR_LOG_FILE_LOCATION}/sa${i} ` | DayOfTheMonth ${i}" | grep -v `hostname` >> ${LOGFILE_PREFIX}_IO.log fi done # Average usage log creation # CPU usage CPU_CORES=`cat /proc/cpuinfo | grep vendor_id | wc -l` echo "CPU Cores : ${CPU_CORES}" > ${LOGFILE_PREFIX}_CPU_AVG.log cat /proc/cpuinfo | grep 'model name' | sort -u >> ${LOGFILE_PREFIX}_CPU_AVG.log cat /proc/cpuinfo | grep 'cpu MHz' | sort -u >> ${LOGFILE_PREFIX}_CPU_AVG.log echo "-----------------------------" >> ${LOGFILE_PREFIX}_CPU_AVG.log echo "`sar -u | sed -n 3p` | Description" >> ${LOGFILE_PREFIX}_CPU_AVG.log cat ${LOGFILE_PREFIX}_CPU.log | grep 'Average:' >> ${LOGFILE_PREFIX}_CPU_AVG.log # RAM usage echo "-----------------------------" > ${LOGFILE_PREFIX}_RAM_AVG.log cat /proc/meminfo | grep 'MemTotal' | sort -u >> ${LOGFILE_PREFIX}_RAM_AVG.log cat /proc/meminfo | grep 'SwapTotal' | sort -u >> ${LOGFILE_PREFIX}_RAM_AVG.log echo "-----------------------------" >> ${LOGFILE_PREFIX}_RAM_AVG.log echo "`sar -r | sed -n 3p` | Description" >> ${LOGFILE_PREFIX}_RAM_AVG.log cat ${LOGFILE_PREFIX}_RAM.log | grep 'Average:' >> ${LOGFILE_PREFIX}_RAM_AVG.log # SWAP usage echo "-----------------------------" > ${LOGFILE_PREFIX}_SWP_AVG.log cat /proc/meminfo | grep 'SwapTotal' | sort -u >> ${LOGFILE_PREFIX}_SWP_AVG.log echo "-----------------------------" >> ${LOGFILE_PREFIX}_SWP_AVG.log echo "`sar -S | sed -n 3p` | Description" >> ${LOGFILE_PREFIX}_SWP_AVG.log cat ${LOGFILE_PREFIX}_SWP.log | grep 'Average:' >> ${LOGFILE_PREFIX}_SWP_AVG.log # Server LOAD echo "-----------------------------" > ${LOGFILE_PREFIX}_LOAD_AVG.log echo "CPU Cores : ${CPU_CORES}" >> ${LOGFILE_PREFIX}_LOAD_AVG.log echo "-----------------------------" >> ${LOGFILE_PREFIX}_LOAD_AVG.log echo "`sar -q | sed -n 3p` | Description" >> ${LOGFILE_PREFIX}_LOAD_AVG.log cat ${LOGFILE_PREFIX}_LOAD.log | grep 'Average:' >> ${LOGFILE_PREFIX}_LOAD_AVG.log # Disk usage echo "------------------------------------------------" > ${LOGFILE_PREFIX}_HDD.log df -Ph | grep -v tmpfs >> ${LOGFILE_PREFIX}_HDD.log echo "------------------------------------------------" >> ${LOGFILE_PREFIX}_HDD.log # I/O activity echo "-----------------------------" > ${LOGFILE_PREFIX}_IO_AVG.log echo "I/O activity on server" >> ${LOGFILE_PREFIX}_IO_AVG.log echo "-----------------------------" >> ${LOGFILE_PREFIX}_IO_AVG.log echo "`sar -b | sed -n 3p` | Description" >> ${LOGFILE_PREFIX}_IO_AVG.log cat ${LOGFILE_PREFIX}_IO.log | grep 'Average:' >> ${LOGFILE_PREFIX}_IO_AVG.log # If you put AVERAGES_USAGE_REPORT=Y you will get this email if [ ${AVERAGES_USAGE_REPORT} == "Y" ] ; then echo "Prosecne vrednosti su u prilogu" | mail -s "${SERVER_NAME} - Average usage report for 30 days" \ -a ${LOGFILE_PREFIX}_CPU_AVG.log \ -a ${LOGFILE_PREFIX}_RAM_AVG.log \ -a ${LOGFILE_PREFIX}_SWP_AVG.log \ -a ${LOGFILE_PREFIX}_LOAD_AVG.log \ -a ${LOGFILE_PREFIX}_IO_AVG.log \ -a ${LOGFILE_PREFIX}_HDD.log \ ${MAIL_TO} fi # If you put DAYLY_BASED_REPORT=Y you will get this email if [ ${DAYLY_BASED_REPORT} == "Y" ] ; then echo "Puna statistika je u prilogu" | mail -s "${SERVER_NAME} - Daily usage report for 30 days" \ -a ${LOGFILE_PREFIX}_CPU.log \ -a ${LOGFILE_PREFIX}_RAM.log \ -a ${LOGFILE_PREFIX}_SWP.log \ -a ${LOGFILE_PREFIX}_LOAD.log \ -a ${LOGFILE_PREFIX}_IO.log \ -a ${LOGFILE_PREFIX}_HDD.log \ ${MAIL_TO} fi exit 0
Here is example of CPU average usage report which you can get to email with this script
