Update monitor-all.sh (#4437)

✅ Summary of Changes from the Initial Version
🧩 1. Tag-Based Filtering (Core Feature)
New feature: Only restart instances (VMs or containers) that have the mon-restart tag.

This makes monitoring and auto-restart controllable directly from the Proxmox Web UI, without editing scripts or services.

Set via GUI: VM → Options → Tags → mon-restart

Set via CLI: qm set <vmid> -tags mon-restart or pct set <ctid> -tags mon-restart

This is the primary new control mechanism, making the script safer, more flexible, and user-friendly.

🧰 2. Backward-Compatible Exclusion Mechanism
The original feature that lets you exclude instances via CLI arguments is preserved:

bash
Copy
Edit
./ping-instances.sh 101 300
These IDs will always be skipped regardless of tag.

🧠 3. Intelligent Responsiveness Checks
For VMs:

Uses qm guest cmd <id> ping to check responsiveness via the QEMU guest agent.

No longer relies on network-level ping, which can be misleading or blocked.

For containers (CTs):

Uses traditional ping to IP addresses obtained from pct exec, since CTs don’t support QEMU agent.

⛔ 4. Instance Skipping Improvements
Instances are now skipped if:

They are explicitly excluded via CLI.

They are templates.

They are configured with onboot: 0 or missing.

They lack the mon-restart tag, regardless of other status.

🪵 5. Same Logging Behavior
All output continues to go to /var/log/ping-instances.log for persistent tracking.

Verbose messages were added for traceability (e.g., why a VM or CT was skipped).

🎯 Why This Matters
With tag-based control, admins can now manage restart behavior dynamically from the Proxmox Web UI, making the script:

More secure (no accidental restarts).

More maintainable (no script edits needed).

More user-friendly (integrated with the UI workflow).

tools/pve/monitor-all.sh

@@ -16,43 +16,52 @@ cat <<"EOF"
 EOF
 
 add() {
+  echo -e "\n IMPORTANT: Tag-Based Monitoring Enabled"
+  echo "Only VMs and containers with the tag 'mon-restart' will be automatically restarted by this service."
+  echo
+  echo "🔧 How to add the tag:"
+  echo "  → Proxmox Web UI: Go to VM/CT → Options → Tags → Add 'mon-restart'"
+  echo "  → CLI: qm set <vmid> -tags mon-restart"
+  echo "         pct set <ctid> -tags mon-restart"
+  echo
+
   while true; do
-    read -p "This script will add Monitor All to Proxmox VE. Proceed(y/n)?" yn
+    read -p "This script will add Monitor All to Proxmox VE. Proceed (y/n)? " yn
     case $yn in
-    [Yy]*) break ;;
+      [Yy]*) break ;;
-    [Nn]*) exit ;;
+      [Nn]*) exit ;;
-    *) echo "Please answer yes or no." ;;
+      *) echo "Please answer yes or no." ;;
     esac
   done
 
-  echo '#!/usr/bin/env bash
+  cat <<'EOF' >/usr/local/bin/ping-instances.sh
+#!/usr/bin/env bash
+
 # Read excluded instances from command line arguments
 excluded_instances=("$@")
 echo "Excluded instances: ${excluded_instances[@]}"
 
 while true; do
 
-  for instance in $(pct list | awk '\''{if(NR>1) print $1}'\''; qm list | awk '\''{if(NR>1) print $1}'\''); do
+  for instance in $(pct list | awk 'NR>1 {print $1}'; qm list | awk 'NR>1 {print $1}'); do
     # Skip excluded instances
     if [[ " ${excluded_instances[@]} " =~ " ${instance} " ]]; then
       echo "Skipping $instance because it is excluded"
       continue
     fi
 
-    # Determine the type of the instance (container or virtual machine)
+    # Determine type and set config command
     if pct status $instance >/dev/null 2>&1; then
-      # It is a container
+      type="ct"
       config_cmd="pct config"
-      IP=$(pct exec $instance ip a s dev eth0 | awk '\''/inet / {print $2}'\'' | cut -d/ -f1)
     else
-      # It is a virtual machine
+      type="vm"
       config_cmd="qm config"
-      IP=$(qm guest cmd $instance network-get-interfaces | egrep -o "([0-9]{1,3}\.){3}[0-9]{1,3}" | grep -E "192\.|10\." | head -n 1)
     fi
 
-    # Skip instances based on onboot and templates
+    # Skip templates and onboot-disabled
     onboot=$($config_cmd $instance | grep -q "onboot: 0" || ( ! $config_cmd $instance | grep -q "onboot" ) && echo "true" || echo "false")
-    template=$($config_cmd $instance | grep template | grep -q "template:" && echo "true" || echo "false")
+    template=$($config_cmd $instance | grep -q "^template:" && echo "true" || echo "false")
 
     if [ "$onboot" == "true" ]; then
       echo "Skipping $instance because it is set not to boot"
@@ -62,36 +71,49 @@ while true; do
       continue
     fi
 
-    # Ping the instance
+    # Check for mon-restart tag
-    if ! ping -c 1 $IP >/dev/null 2>&1; then
+    has_tag=$($config_cmd $instance | grep -q "tags:.*mon-restart" && echo "true" || echo "false")
-      # If the instance can not be pinged, stop and start it
+    if [ "$has_tag" != "true" ]; then
-      if pct status $instance >/dev/null 2>&1; then
+      echo "Skipping $instance because it does not have 'mon-restart' tag"
-        # It is a container
+      continue
+    fi
+
+    # Responsiveness check and restart if needed
+    if [ "$type" == "vm" ]; then
+      # Check if guest agent responds
+      if qm guest cmd $instance ping >/dev/null 2>&1; then
+        echo "VM $instance is responsive via guest agent"
+      else
+        echo "$(date): VM $instance is not responding to agent ping, restarting..."
+        if qm status $instance | grep -q "status: running"; then
+          qm stop $instance >/dev/null 2>&1
+          sleep 5
+        fi
+        qm start $instance >/dev/null 2>&1
+      fi
+    else
+      # Container: get IP and ping
+      IP=$(pct exec $instance ip a s dev eth0 | awk '/inet / {print $2}' | cut -d/ -f1)
+      if ! ping -c 1 $IP >/dev/null 2>&1; then
         echo "$(date): CT $instance is not responding, restarting..."
         pct stop $instance >/dev/null 2>&1
         sleep 5
         pct start $instance >/dev/null 2>&1
       else
-        # It is a virtual machine
+        echo "CT $instance is responsive"
-        if qm status $instance | grep -q "status: running"; then
-          echo "$(date): VM $instance is not responding, restarting..."
-          qm stop $instance >/dev/null 2>&1
-          sleep 5
-        else
-          echo "$(date): VM $instance is not running, starting..."
-        fi
-        qm start $instance >/dev/null 2>&1
       fi
     fi
   done
 
-  # Wait for 5 minutes. (Edit to your needs)
   echo "$(date): Pausing for 5 minutes..."
   sleep 300
-done >/var/log/ping-instances.log 2>&1' >/usr/local/bin/ping-instances.sh
+
+done >/var/log/ping-instances.log 2>&1
+EOF
+
   touch /var/log/ping-instances.log
-  # Change file permissions to executable
   chmod +x /usr/local/bin/ping-instances.sh
+
   cat <<EOF >/etc/systemd/system/ping-instances.timer
 [Unit]
 Description=Delay ping-instances.service by 5 minutes
@@ -104,17 +126,17 @@ OnUnitActiveSec=300
 WantedBy=timers.target
 EOF
 
-  # Create ping-instances.service
   cat <<EOF >/etc/systemd/system/ping-instances.service
 [Unit]
-Description=Ping instances every 5 minutes and restarts if necessary
+Description=Ping instances every 5 minutes and restart if necessary
 After=ping-instances.timer
 Requires=ping-instances.timer
+
 [Service]
 Type=simple
-# To specify which CT/VM should be excluded, add the CT/VM ID at the end of the line where ExecStart=/usr/local/bin/ping-instances.sh is specified.
+# To exclude specific instances, pass IDs to ExecStart, e.g.:
-# For example: ExecStart=/usr/local/bin/ping-instances.sh 100 102
+# ExecStart=/usr/local/bin/ping-instances.sh 100 200
-# Virtual machines without the QEMU guest agent installed must be excluded.
+# Instances must also have the 'mon-restart' tag to be monitored
 
 ExecStart=/usr/local/bin/ping-instances.sh
 Restart=always
@@ -125,39 +147,33 @@ StandardError=file:/var/log/ping-instances.log
 WantedBy=multi-user.target
 EOF
 
-  # Reload daemon, enable and start ping-instances.service
   systemctl daemon-reload
   systemctl enable -q --now ping-instances.timer
   systemctl enable -q --now ping-instances.service
   clear
-  echo -e "\n To view Monitor All logs: cat /var/log/ping-instances.log"
+  echo -e "\n Monitor All installed."
+  echo "📄 To view logs: cat /var/log/ping-instances.log"
+  echo "⚙️  Make sure your VMs or containers have the 'mon-restart' tag to be monitored."
 }
 
 remove() {
   systemctl disable -q --now ping-instances.timer
   systemctl disable -q --now ping-instances.service
-  rm /etc/systemd/system/ping-instances.service /etc/systemd/system/ping-instances.timer /usr/local/bin/ping-instances.sh /var/log/ping-instances.log
+  rm -f /etc/systemd/system/ping-instances.service
-  echo "Removed Monitor All from Proxmox VE"
+  rm -f /etc/systemd/system/ping-instances.timer
+  rm -f /usr/local/bin/ping-instances.sh
+  rm -f /var/log/ping-instances.log
+  echo "Monitor All removed from Proxmox VE"
 }
 
-# Define options for the whiptail menu
 OPTIONS=(Add "Add Monitor-All to Proxmox VE"
   Remove "Remove Monitor-All from Proxmox VE")
 
-# Show the whiptail menu and save the user's choice
 CHOICE=$(whiptail --backtitle "Proxmox VE Helper Scripts" --title "Monitor-All for Proxmox VE" --menu "Select an option:" 10 58 2 \
   "${OPTIONS[@]}" 3>&1 1>&2 2>&3)
 
-# Check the user's choice and perform the corresponding action
 case $CHOICE in
-"Add")
+"Add") add ;;
-  add
+"Remove") remove ;;
-  ;;
+*) echo "Exiting..."; exit 0 ;;
-"Remove")
-  remove
-  ;;
-*)
-  echo "Exiting..."
-  exit 0
-  ;;
 esac