Monit service monitoring - Alerts - Web UI - Enhance Control Panel

Monit service monitoring - Alerts - Web UI

gmakhs

Monitoring services on your server is crucial for ensuring reliability and rapid recovery from downtime.
Monit is a simple, lightweight tool perfect for monitoring and managing services automatically.
I personally use it and adjust this general config file to monitor the service i want per server, is lite weight and works with enhance.

How to Setup

Step 1: Install Monit

For Debian/Ubuntu:

sudo apt update
sudo apt install monit

Step 2: Basic Monit Configuration

Edit Monit's config file:

sudo nano /etc/monit/monitrc

Enable the web interface (optional):

set httpd port 2812 and
    use address 0.0.0.0
    allow admin:monit

Step 3: Configuring Monit to Monitor Essential Services

Add the following to /etc/monit/monitrc:

#SSH

check program sshd with path "/bin/systemctl is-active --quiet ssh"
    if status != 0 then exec "/bin/systemctl restart ssh"
    if 5 restarts within 5 cycles then timeout

#Cron

check program cron with path "/bin/systemctl is-active --quiet cron"
    if status != 0 then exec "/bin/systemctl restart cron"
    if 5 restarts within 5 cycles then timeout

#Rspamd

check program rspamd with path "/bin/systemctl is-active --quiet rspamd"
    if status != 0 then exec "/bin/systemctl restart rspamd"
    if 5 restarts within 5 cycles then timeout

#Postfix

check program postfix with path "/bin/systemctl is-active --quiet postfix"
    if status != 0 then exec "/bin/systemctl restart postfix"
    if 5 restarts within 5 cycles then timeout

#Dovecot

check program dovecot with path "/bin/systemctl is-active --quiet dovecot"
    if status != 0 then exec "/bin/systemctl restart dovecot"
    if 5 restarts within 5 cycles then timeout

#Pure-FTP

check program pure-ftpd with path "/bin/systemctl is-active --quiet pure-ftpd"
    if status != 0 then exec "/bin/systemctl restart pure-ftpd"
    if 5 restarts within 5 cycles then timeout

#LiteSpeed Web Server (LSWS)

check program lsws with path "/bin/systemctl is-active --quiet lsws"
    if status != 0 then exec "/bin/systemctl restart lsws"
    if 5 restarts within 5 cycles then timeout

#Nginx or Apache (choose one and uncomment)

# Nginx
#check program nginx with path "/bin/systemctl is-active --quiet nginx"
#    if status != 0 then exec "/bin/systemctl restart nginx"
#    if 5 restarts within 5 cycles then timeout

# Apache
#check program apache2 with path "/bin/systemctl is-active --quiet apache2"
#    if status != 0 then exec "/bin/systemctl restart apache2"
#    if 5 restarts within 5 cycles then timeout


#Server load
check system localhost
    if loadavg (1min) > 8 then alert

Step 5: Verify Status

Check status easily with:

sudo monit status

Alerts

Set up email alerts in your monitrc to receive notifications for service restarts and failures:

set mailserver smtp.example.com port 587
    username "user@example.com" password "yourpassword"
    using tls

set alert admin@example.com

Additional Examples

# Alert on high RAM usage
#check system localhost
#    if memory usage > 90% then alert

# Alert on disk usage
#check filesystem rootfs with path /
#    if space usage > 80% then alert

# Alert on inode usage
#check filesystem rootfs with path /
#    if inode usage > 90% then alert

Now Monit will automatically monitor, restart services as needed, and alert you via email if anything goes wrong, ensuring your server remains stable and reliable

Vendoz

I didn't know about monit but it looks really nice! Light, simple and might do the job, I'll try it, thanks!

gmakhs

Vendoz I didn't know also until recently, but it has become very handy !

gmakhs

It can also send Alerts for resource usage of websites like :

`` Description: status failed (1) -- UUID USED% USED(MB) LIMIT(MB)
41027b78-4eb6-41ad-8473-8171cf12dd20 77% 1542MB 2000MB
0e9c45da-32f5-4cba-8d25-37c677d479bb 75% 1509MB 2000MB
82437c65-7dd7-4afe-aa0c-9f519ff72f8e 65% 1350MB 2048MB
84395b38-dee1-4aa1-9858-5476c47c0b18 38% 393MB 1024MB
215bc9af-54f1-4de9-8e96-88434caad2cd 34% 350MB 1024MB
551e419f-22a6-4e9f-9121-cdfc1a6669fa 25% 531MB 2048MB
8d9bd017-704f-4ed4-83e6-c3c301eba8be 22% 228MB 1024MB
99b42f4a-d152-4a60-82d2-cfb88507bd41 18% 368MB 2048MB
5130908e-c84a-41f7-9df1-136b5fe6ce15 14% 145MB 1024MB
3e6b54b0-404f-4a75-8807-84cf6e4b2a69 13% 140MB

``
i am gonna make a script and publish it

prasad0889

Thanks @gmakhs, Looking forward to your script.

gmakhs

prasad0889 this is how it looks :

https://imgur.com/a/6fa4v5B](https://imgur.com/a/6fa4v5B
This is the script :

#!/bin/bash

MODE=${1:-cpu}

MEMORY_LIMIT_PERCENT=90
CPU_LIMIT_PERCENT=90
TASK_LIMIT=90
NPROC_LIMIT=90

USER="www-data"

ALERT=0
MSG=""

CPUS=$(nproc)

TMP_FILE_BEFORE=$(mktemp)
TMP_FILE_AFTER=$(mktemp)

# Save CPU usage at T1
for UUID_PATH in /sys/fs/cgroup/websites/*; do
  UUID=$(basename "$UUID_PATH")
  if [ -f "$UUID_PATH/cpu.stat" ]; then
    awk '/usage_usec/ {print $2}' "$UUID_PATH/cpu.stat" | xargs echo "$UUID" >> "$TMP_FILE_BEFORE"
  fi
done

sleep 10

# Save CPU usage at T2
for UUID_PATH in /sys/fs/cgroup/websites/*; do
  UUID=$(basename "$UUID_PATH")
  if [ -f "$UUID_PATH/cpu.stat" ]; then
    awk '/usage_usec/ {print $2}' "$UUID_PATH/cpu.stat" | xargs echo "$UUID" >> "$TMP_FILE_AFTER"
  fi
done

TMP_FILE=$(mktemp)

while read -r line1 && read -r line2 <&3; do
  UUID1=$(echo "$line1" | awk '{print $1}')
  CPU1=$(echo "$line1" | awk '{print $2}')
  
  UUID2=$(echo "$line2" | awk '{print $1}')
  CPU2=$(echo "$line2" | awk '{print $2}')
  
  [ "$UUID1" != "$UUID2" ] && continue
  
  CPU_DELTA=$((CPU2 - CPU1))
  CPU_PERCENT=$((CPU_DELTA * 100 / (1000000 * CPUS)))

  UUID_PATH="/sys/fs/cgroup/websites/$UUID1"

  # Memory
  if [ -f "$UUID_PATH/memory.current" ]; then
    MEM=$(cat "$UUID_PATH/memory.current")
    LIMIT=$(cat "$UUID_PATH/memory.max")
    [ "$LIMIT" = "max" ] && LIMIT=$((128 * 1024 * 1024 * 1024))
    MEM_PERCENT=$((MEM * 100 / LIMIT))
    MEM_MB=$((MEM / 1024 / 1024))
    LIMIT_MB=$((LIMIT / 1024 / 1024))
  fi

  # Tasks
  [ -f "$UUID_PATH/pids.current" ] && TASKS=$(cat "$UUID_PATH/pids.current")

  # NPROC
  NPROC=$(ps -u "$USER" --no-headers | wc -l)

  # Save to tmp
  echo "$UUID1 $CPU_PERCENT $MEM_PERCENT $MEM_MB $LIMIT_MB $TASKS $NPROC" >> "$TMP_FILE"

  # Check limits
  [ "$MEM_PERCENT" -gt "$MEMORY_LIMIT_PERCENT" ] && MSG="${MSG}High Memory: ${MEM_PERCENT}% used by $UUID1\n" && ALERT=1
  [ "$CPU_PERCENT" -gt "$CPU_LIMIT_PERCENT" ] && MSG="${MSG}High CPU: ${CPU_PERCENT}% used by $UUID1\n" && ALERT=1
  [ "$TASKS" -gt "$TASK_LIMIT" ] && MSG="${MSG}High Tasks: ${TASKS} running in $UUID1\n" && ALERT=1
  [ "$NPROC" -gt "$NPROC_LIMIT" ] && MSG="${MSG}High NPROC: ${NPROC} processes for $USER\n" && ALERT=1
done <"$TMP_FILE_BEFORE" 3<"$TMP_FILE_AFTER"

# Pretty display Top 10
printf "%-40s %-7s %-7s %-10s %-10s %-7s %-7s\n" "UUID" "CPU%" "MEM%" "USED(MB)" "LIMIT(MB)" "TASKS" "NPROC"

case "$MODE" in
  cpu)
    sort -k2 -rn "$TMP_FILE" | head -n 10 | awk '{printf "%-40s %-7s %-7s %-10s %-10s %-7s %-7s\n", $1, $2"%", $3"%", $4"MB", $5"MB", $6, $7}'
    ;;
  ram|mem)
    sort -k3 -rn "$TMP_FILE" | head -n 10 | awk '{printf "%-40s %-7s %-7s %-10s %-10s %-7s %-7s\n", $1, $2"%", $3"%", $4"MB", $5"MB", $6, $7}'
    ;;
  nproc)
    sort -k7 -rn "$TMP_FILE" | head -n 10 | awk '{printf "%-40s %-7s %-7s %-10s %-10s %-7s %-7s\n", $1, $2"%", $3"%", $4"MB", $5"MB", $6, $7}'
    ;;
  *)
    echo "Invalid mode: $MODE. Use: cpu | ram | mem | nproc"
    ;;
esac

# Cleanup
rm -f "$TMP_FILE" "$TMP_FILE_BEFORE" "$TMP_FILE_AFTER"

# Final check
if [ "$ALERT" -eq 1 ]; then
  echo -e "$MSG"
  exit 1
else
  exit 0
fi

i run it in monit us

check program cgroup_usage with path "/bin/bash /usr/local/bin/cgroup_check_all.sh"
#  if status != 0 for 5 cycles then alert
  if status != 0 then alert

I am trying to also get disk usage

standtech

gmakhs There is a command you can use "systemd-cgtop" eg. "systemd-cgtop websites -m"

gmakhs

standtech yes but this command won't send an email warning 🙂

standtech

gmakhs I was just sharing a command that could be helpful for people reading this thread, but yes, this community is hands-on and expert. Regarding email warnings, if you can set quotas on website resources, then monitoring their usage is overhead but again depends on use cases. Just my opinion 😉

gmakhs

standtech of course !!!
I am working now on quotas andxa plugin for blesta to read the results and warn users to upgrade if certain criteria's are meet