System Resources Monitoring

System Resources Monitoring

Version Info

RouterOS 7.16 | Min: v6.0+ | ⏱️ 5 minutes | 🟢 Beginner

Tested on: 7.16 7.14

Version-specific notes

• v7.0: Added RPS (Receive Packet Steering) submenu
• v6.x: Same commands work on RouterOS 6.x

TL;DR (Quick Start)

For the impatient: view system resources and CPU usage.

/system resource print

For per-process CPU breakdown:

/tool profile

Overview

What this does: System Resources provides real-time monitoring of CPU load, memory usage, disk space, uptime, and hardware information. The profiler tool identifies which processes consume CPU resources.

When to use this:

• Monitoring router health and performance
• Troubleshooting high CPU or memory usage
• Verifying hardware specifications
• Capacity planning before adding services
• Diagnosing slow network performance

Prerequisites:

• Access to RouterOS via Winbox, SSH, or WebFig
• Read permissions (no special access required)

Configuration Steps

Step 1: View System Resources

Display overall system information and resource usage.

/system resource print

Expected output:

                   uptime: 3d14h52m31s
                  version: 7.16 (stable)
               build-time: 2024-09-20 13:29:57
         factory-software: 7.10
              free-memory: 186.5MiB
             total-memory: 256.0MiB
                      cpu: ARM
                cpu-count: 2
            cpu-frequency: 800MHz
                 cpu-load: 3%
           free-hdd-space: 12.4MiB
          total-hdd-space: 16.0MiB
  write-sect-since-reboot: 4521
         write-sect-total: 124589
               bad-blocks: 0%
        architecture-name: arm
               board-name: hAP ax lite
                 platform: MikroTik

Step 2: Monitor Real-Time Usage

View live CPU and memory statistics.

/system resource monitor once

Expected output:

       cpu-used: 5%
cpu-used-per-cpu: 3%,7%
    free-memory: 186.0MiB

For continuous monitoring (press Ctrl+C to stop):

/system resource monitor

Step 3: View Per-Core CPU Usage

Check individual CPU core loads.

/system resource cpu print

Expected output:

 # CPU  LOAD  IRQ  DISK
 0   0    2%   1%    0%
 1   1    4%   0%    0%

Step 4: Identify CPU-Heavy Processes

Use the profiler to find which processes consume CPU.

/tool profile

Expected output:

NAME              CPU     USAGE
idle              all     97%
networking        all     2%
management        all     1%

For per-core breakdown:

/tool profile cpu=all

Common Scenarios

Scenario: Troubleshoot High CPU Usage

When CPU load is consistently high, identify the cause:

# Check overall CPU
/system resource print

# Identify process consuming CPU
/tool profile duration=10

Common high-CPU processes:

Process	Likely Cause
networking	High packet rate, consider hardware offload
management	Too many Winbox/SSH sessions, scripts running
firewall	Complex firewall rules, consider fasttrack
dns	DNS cache under heavy load
wireless	Many wireless clients, interference
encrypting	VPN/IPsec traffic without hardware acceleration

Scenario: Monitor Memory Usage

Check for memory pressure:

/system resource print

Calculate usage percentage:

:local free [/system resource get free-memory]
:local total [/system resource get total-memory]
:put ("Memory used: " . (($total - $free) * 100 / $total) . "%")

Memory guidelines:

• Below 20% free: Consider reducing services or upgrading hardware
• Below 10% free: Risk of instability, immediate action needed
• Out of memory: Router may reboot unexpectedly

Scenario: Check Disk Space

Monitor storage for logs and files:

/system resource print

If disk is full:

# Check file usage
/file print

# Clear old logs if needed
/system logging action set memory memory-lines=100

Scenario: View Hardware Information

List connected hardware devices:

/system resource hardware print

For USB devices:

/system resource usb print

Scenario: Script-Based Monitoring

Get specific values for scripts or SNMP:

# Get free memory
/system resource get free-memory

# Get CPU load
/system resource get cpu-load

# Get uptime
/system resource get uptime

Example alert script:

:local cpuLoad [/system resource get cpu-load]
:if ($cpuLoad > 80) do={
    :log warning ("High CPU: " . $cpuLoad . "%")
}

Scenario: IRQ Load Balancing (Multi-Core)

On multi-core systems with high network traffic, distribute IRQ load:

# View current IRQ assignments
/system resource irq print

# Check per-core IRQ load
/system resource cpu print

Enable RPS for better packet distribution (v7+):

/system resource irq rps set [find] rps=yes

Verification

Check 1: Verify Resources are Accessible

/system resource print

Expected: Shows uptime, version, memory, and CPU information.

Check 2: Verify CPU is Not Overloaded

/system resource get cpu-load

Expected: Below 80% under normal operation.

Check 3: Verify Sufficient Free Memory

/system resource get free-memory

Expected: At least 20% of total memory free.

Troubleshooting

Symptom	Cause	Solution
CPU constantly at 100%	Overloaded or attack	Use /tool profile to identify process; enable fasttrack for routing
Memory slowly decreasing	Memory leak or growing tables	Check connection tracking, DNS cache, address lists; consider reboot
”Out of memory” reboot	Insufficient RAM for workload	Reduce services, clear logs, disable unused packages, upgrade hardware
High IRQ usage	Network traffic on single core	Enable hardware offload; use RPS on v7+
Disk space full	Log files or packages	Clear files with /file remove; reduce logging verbosity
management high CPU	Many concurrent sessions	Limit Winbox sessions; check for runaway scripts
Profiler shows unclassified	Unknown process	Usually kernel tasks; monitor for patterns

Common Mistakes

• Don’t ignore gradual memory decline - Memory that never returns may indicate a leak; plan periodic reboots if needed
• Don’t confuse profile total with resource load - /tool profile percentages may not match /system resource exactly
• Don’t overlook IRQ distribution - Single-core IRQ handling limits throughput on multi-core routers
• Don’t forget disk writes wear flash - Excessive logging shortens NAND lifespan; use remote syslog for busy routers

Community Tips

Memory Usage Mystery: RouterOS doesn’t show per-process memory usage. If memory is low, check: connection tracking entries, DNS cache size, firewall address lists, and log buffer.

CPU Profiler Accuracy: Run profiler for at least 10 seconds (duration=10) for accurate readings. Brief samples may miss intermittent spikes.

Fasttrack for Performance: Enable fasttrack to bypass firewall for established connections:

/ip firewall filter add chain=forward action=fasttrack-connection connection-state=established,related

Profiler Process Reference

Process	Description
idle	Free CPU (higher is better)
networking	Packet processing
management	Winbox, SSH, API, web sessions
firewall	Packet filtering and NAT
encrypting	IPsec, VPN encryption
dns	DNS server/cache
routing	Routing protocol calculations
wireless	Wireless driver operations
bridging	Bridge forwarding
unclassified	Kernel and uncategorized tasks

Related Topics

Monitoring

• Health Monitoring - hardware sensors and voltage
• Graphing - enable resource usage graphs
• Netwatch - monitor hosts and trigger alerts
• Logging - configure system logging

Performance

• Profiler - identify CPU-intensive processes
• Torch - real-time traffic analysis
• Scheduler - automate resource monitoring

System Management

• System Backup - backup before changes
• RouterBOARD - hardware information
• Package Installation - manage features

Reference

• MikroTik Resource Documentation
• MikroTik Profiler Documentation

Properties Reference

Property	Type	Description
uptime	time	Time since last boot
version	string	RouterOS version
cpu-load	percent	Current CPU usage
free-memory	bytes	Available RAM
total-memory	bytes	Installed RAM
free-hdd-space	bytes	Available storage
total-hdd-space	bytes	Total storage
cpu-count	integer	Number of CPU cores
cpu-frequency	MHz	CPU clock speed
architecture-name	string	CPU architecture (arm, x86, etc.)
board-name	string	Hardware model
bad-blocks	percent	NAND bad block percentage

Command Reference

Command	Description
/system resource print	Display all resource information
/system resource monitor	Real-time CPU and memory monitoring
/system resource cpu print	Per-core CPU statistics
/system resource irq print	Interrupt assignments
/system resource hardware print	Connected hardware devices
/system resource usb print	USB device information
/system resource get <property>	Get specific value for scripting
/tool profile	Per-process CPU profiler
/tool profile cpu=all	Per-core process breakdown

Summary

System Resources monitoring is essential for router health and troubleshooting:

1. View resources with /system resource print for CPU, memory, disk, and uptime
2. Monitor real-time with /system resource monitor for live updates
3. Identify CPU hogs with /tool profile to see per-process usage
4. Check per-core with /system resource cpu print for load distribution

Key points:

• CPU load below 80% is healthy; above indicates overload or attack
• Keep at least 20% memory free to avoid instability
• Use /tool profile duration=10 for accurate CPU breakdown (brief samples miss spikes)
• High networking in profiler suggests enabling fasttrack or hardware offload
• Multi-core systems benefit from RPS (Receive Packet Steering) in v7+
• Excessive disk writes wear flash storage; use remote syslog for busy routers
• Memory usage per-process isn’t shown; check connection tracking, DNS cache, and address lists
• Gradual memory decline may indicate a leak; plan periodic maintenance reboots if needed