Server Performance Stats

Here i have my notes, errors and thought process it’s just the notes i took in order to acomplish this project, what i wanted to do is a script that gives metrics on CPU, disk, memory, processes and also useful info like failed logs and users logged in to see the script itself check The Github repo

the shebang #! is to specify to Linux to run this as a bash script specifying the shell

#!/bin/bash
# server-stats.sh
# Author: Mohamed Kamil El Kouarti
# Last update: Thu Sep 25 2025

# Will call later in the os name and version part
 . /etc/os-release

Then i created color variables that i will use afterwards so the output has better readability the \033 tells to bash t interpret it as a formatting instruction not as a normal text and what follows is the style and color in ANSI

    # Color codes
    RED='\033[0;31m'
    GREEN='\033[0;32m'
    YELLOW='\033[1;33m'
    BLUE='\033[0;34m'
    PURPLE='\033[0;35m'
    CYAN='\033[0;36m'
    WHITE='\033[1;37m'
    NC='\033[0m'

I created the variable to get the used memory with free -m to get it on megabytes and then i added a pipe so i can filter the output of that command with awk the NR==2 to get to the second line, and then i sued {print $2} to get the second column i will be using the same method in the following to get the other stats.

#MEMORY USAGE
used=$(free -m | awk 'NR==2 { print $3 }')

total=$(free -m | awk 'NR==2 { print $2 }')

free_mem=$(free -m | awk 'NR==2 {print $4}')

Then to get the disk usage i first put LC_NUMERIC=C to avoid using the coma , as a decimal separator which makes the script more universal and consistent then the df -B1 to get the disk stats in bytes so i can later manuallyconvert it to MG, then as i did before use NR==2 to get to the second line, used printf instead of print to get a formatted output in this case 2 decimals with “%.2f” and lastly specified the column with $number and the number on that column will be converted to GB

#DISK USAGE
free_disk=$(LC_NUMERIC=C df -B1 / | awk 'NR==2 {printf "%.2f", $4/1024/1024/1024}')
used_disk=$(LC_NUMERIC=C df -B1 / | awk 'NR==2 {printf "%.2f", $3/1024/1024/1024}')
percentage=$(df -h / | awk 'NR==2 {print $5}')
total_disk=$(LC_NUMERIC=C df -B1 / | awk 'NR==2 {printf "%.2f", $2/1024/1024/1024}')

To get the cu usage i used the top -b -1 to get an static output then i used ‘/%Cpu/ to get the line containing %Cpu for(i=1;<=NF;i++) i=1 means current position and examines each word in the Cpu line in search of , until the last character i<==N and i++ moves onto the next word in loop if the $i (current word) equals the word id then print formated (put 2 decimals with a . ) and then 100 -$(i-1) to get the percentage

# CPU USAGE 
total_cpu=$(top -b -n1 | awk '/%Cpu/ {for(i=1;i<=NF;i++) if($i=="id,") {printf "%.2f", 100 - $(i-1)}}')

To get the top 5 cpu and top 5 memory i used the ps aux to get the processes and added –sort=-%cpu to get them ordered from highest to lowest with the ‘-‘ then print th 2 3 11 columns to get their PID, CPU USAGE and NAME

top_5_cpu=$(ps aux --sort=-%cpu | awk 'NR>1 && NR<=6 {print $2, $3, $11}')

top_5_mem=$(ps aux --sort=-%mem | awk 'NR>1 && NR<=6 {print $2, $4, $11}')  

To get the OS name and version at first i used the lsb-release but later changed it because some minimalist distros like Alpine dont include it so i switched to /etc/os-release since currently is the most common archive. If /etc/os-release exists (with the -f) enters the ‘then’ and calls the source with . path/to/file which is the equivalent of calling a lib in python. In that file NAME and VERSION contained the values of the device so i just call those variables if the file doesnt exist it just gives them the “unknown value”

if [ -f /etc/os-release ]; then
    os_name="$NAME"
    os_version="$VERSION"
else
    os_name="Unknown"
    os_version="Unknown"
fi

It’s just the regular uptime command but with the -p added to make it easier to read

uptime=$(uptime -p)

To get the users logged i used the who command then i get the first column which contains the usernames with sort -u i ensure to not get duplikates, and the tr ‘\n’ ‘ ‘ to changes all line jumps into spaces so it fits the output better.

users=$(who | awk '{print $1}' | sort -u | tr '\n' ' ')  # Gets unique users

To get the load average then awk -F’load average’ so it takes the load average as a column separator and then the other elemnts are counted as columns. and the sed part is to remove all the spaces at the beginning so it’s by swaping ^means beginning * 0 or more spaces for nothing //.

load_average=$(uptime | awk -F'load average:' '{print $2}' | sed 's/^ *//')  

To get the failed logs i did the next: check each file in $log_files, if $log is a file and exists then the grep is to search thourght log, the -i is to not differentiate between uppercase and lowercase then serach for pam_unix.*fail with * meaning anything in between 2>/dev/null redirects any error message (which is stderr and has the code 2) to /dev/null , a sort of trash if $log has not files it prints the content

log_files="/var/log/secure /var/log/auth.log /var/log/messages /var/log/syslog"

failed_logs_j=$(
for log in $log_files; do
    if [ -f "$log" ]; then
        grep -i "pam_unix.*fail" "$log" 2>/dev/null | head -5
    else
        echo "file not found: $log"
    fi
done
)

To get the memory usage to put a percentage i used awk BEGIN which is useful to do mathematical operations with AWK printf is to formatted print to get 2 decimals and finally to get the percentage it divides the used memory with the total and multiplies it with 100 to get percentage.

memory_usage=$(awk "BEGIN {printf \"%.2f\", $used / $total * 100}")

And lastly i added the colors and variables

echo -e "$(cat <<EOF
${CYAN}=== Server Performance Stats ===${NC}
${BLUE}Last update: $(date)${NC}

${YELLOW}=== CPU Usage ===${NC}
Total CPU Usage: $total_cpu %

${YELLOW}=== Memory Usage ===${NC}
Total Memory: $total MB
Used Memory: $used MB  
Free Memory: $free_mem MB
Memory Usage: $memory_usage %

${YELLOW}=== Disk Usage ===${NC}
Total Disk: $total_disk GB
Used Disk: $used_disk GB
Free Disk: $free_disk GB
Disk Usage: $percentage %

${YELLOW}=== Top 5 Processes by CPU Usage ===${NC}
$top_5_cpu

${YELLOW}=== Top 5 Processes by Memory Usage ===${NC}
$top_5_mem

${YELLOW}=== Additional System Information ===${NC}
OS : $os_name $os_version
Uptime: $uptime
Load Average: $load_average
Logged-in Users: $users

${RED}=== Recent Failed Login Attempts ===${NC}
$failed_logs_j

${PURPLE}=== End of Report ===${NC}
EOF
)"