CPU USAGE

top -bn1 | grep "Cpu(s)" | awk '{print "CPU: " $2 "%"}'

    top -bn1
│
├── top    → shows running processes and system stats
├── -b     → batch mode — runs top non-interactively
│            without -b, top opens as live screen
│            with -b, top prints output and exits
└── -n1    → run only 1 iteration then stop
             without -n1, top keeps refreshing forever

output looks like:
%Cpu(s):  5.2 us,  1.3 sy,  0.0 ni, 92.8 id

| grep "Cpu(s)"
│
└── filters only the CPU line from top output
    ignores all other lines like process list, memory etc
    we only want the line that has CPU stats

| awk '{print "CPU: " $2 "%"}'
│
├── awk    → text processing tool, splits line into fields
├── $2     → second field of that line
│            %Cpu(s):  5.2 us ...
│            $1 = %Cpu(s):
│            $2 = 5.2      ← this is user CPU usage
│            $3 = us
└── print  → prints "CPU: 5.2%"

MEMORY USAGE

free -h | awk 'NR==2 { print "Total: " $2 "Used: " $3 "Free: " $4 "Usage: " $3/$2*100 "%"}'

free -h
│
├── free   → shows memory statistics
└── -h     → human readable format (MB, GB instead of bytes)

output looks like:
              total    used    free
Mem:          7.7Gi   3.2Gi   4.5Gi
Swap:         2.0Gi   0.0Gi   2.0Gi

| awk 'NR==2 {print "Used: " $3 " / Total: " $2}'
│
├── NR==2  → NR = Number of Row
│            NR==2 means process only row 2
│            row 1 = header (total used free)
│            row 2 = Mem:  ← we want this
│            row 3 = Swap: ← skip this
│
├── $2     → second field = total memory  (7.7Gi)
├── $3     → third field  = used memory   (3.2Gi)
└── print  → prints "Used: 3.2Gi / Total: 7.7Gi"

Check-server

In this script if the servers ip is wrong at this point it keeps waiting and wont proceed with checking the other servers. Hence we have added the connection timeout

ssh ec2-user@"$server" "df -hT"
ssh -o ConnectTimeout=5 ec2-user@"$server" "df -hT"

Logrotate

/etc/logrotate.d/myapp

/var/log/myapp/*.log {
    daily
    rotate 7
    compress
    missingok
    notifempty
    copytruncate
}

copytruncate: Copies current log Empties original log file

logrotate -d /etc/logrotate.d/myapp -debug mode
logrotate -f /etc/logrotate.d/myapp -force rotation

After log rotation and compression, we can upload old .gz log files to Amazon S3 using AWS CLI.

#!/bin/bash

LOG_PATH="/var/log/myapp"
S3_BUCKET="s3://my-log-backup-bucket"

find $LOG_PATH -name "*.gz" -mtime +1 -exec aws s3 mv {} $S3_BUCKET \;

we can define cronjob for this

0 1 * * * /opt/scripts/log_backup.sh

Runs daily at 1 AM