Monitoring and Observability

ML Defender provides comprehensive monitoring capabilities through logs, metrics, IPSet statistics, and real-time dashboards. This guide covers log locations, monitoring commands, and observability tools.

Log Locations

Vagrant/Development

All logs are centralized in /vagrant/logs/lab/:

/vagrant/logs/lab/
├── firewall-agent.log      # Firewall ACL Agent
├── firewall-metrics.json   # Firewall metrics export
├── detector.log            # ML Detector
├── sniffer.log             # eBPF Sniffer
├── etcd-server.log         # etcd configuration server
└── rag.log                 # RAG Security System

Docker Compose

# View logs with docker-compose
docker-compose logs -f service1      # Sniffer
docker-compose logs -f service2      # Detector
docker-compose logs --tail=100 etcd  # etcd

# Logs are also written to bind-mounted volumes
./logs/
├── service1.log
├── service2.log
└── etcd.log

Debian Package

# Systemd journal
sudo journalctl -u sniffer-ebpf -f
sudo journalctl -u sniffer-ebpf --since "10 minutes ago"

# Application logs
/var/log/ml-defender/
├── sniffer.log
├── detector.log
└── firewall.log

Monitoring Dashboard

Use the built-in monitoring script for real-time visibility.

Live Monitoring Script

The monitor_lab.sh script provides a comprehensive real-time dashboard:

# Start live monitoring (auto-refreshes every 3 seconds)
cd /vagrant
bash scripts/monitor_lab.sh

# Or use alias (Vagrant)
logs-lab

Dashboard Output:

╔════════════════════════════════════════════════════════════╗
║  ML Defender Lab - Live Monitoring (Enhanced v2.3)         ║
║  2026-02-08 14:32:45                                       ║
║  System Uptime: 2 hours 15 minutes                         ║
╚════════════════════════════════════════════════════════════╝

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📈 System Statistics
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
CPU: 42%
RAM: 56%
Disk: 38%

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📊 Component Status & Configuration
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔥 Firewall:  ✅ PID 12345 - CPU: 5.2% MEM: 1.8% (127MB) - Uptime: 2h 15m
   Config: firewall.json

🤖 Detector:  ✅ PID 12346 - CPU: 8.5% MEM: 3.2% (256MB) - Uptime: 2h 15m
   Config: ml_detector_config.json

📡 Sniffer:   ✅ PID 12347 - CPU: 12.1% MEM: 2.4% (189MB) - Uptime: 2h 14m
  Profile: dual_nic_gateway
  Interface: eth3

🗄️  etcd-server: ✅ PID 12348 - CPU: 1.2% MEM: 0.8% (64MB) - Uptime: 2h 15m
   Status: ✅ Healthy
          Config: etcd.conf

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔌 Communication Channels
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Port 5571 (Sniffer → Detector): ✅ Listening (1 connections)
Port 5572 (Detector → Firewall): ✅ Listening (1 connections)
Port 2379 (etcd client): ✅ Listening (3 connections)
Port 2380 (etcd peer): ✅ Listening (0 connections)

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔥 IPSet Blacklist Status
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
ml_defender_blacklist: ✅ Active - Entries: 42 - Memory: 2384B

Recent blocked IPs:
  • 192.168.1.100
  • 10.0.0.50
  • 172.16.1.200
  • 203.0.113.45
  • 198.51.100.123

Dashboard Features

The monitoring script shows:

System Stats: CPU, RAM, Disk usage
Component Status: Process health, PID, resource usage, uptime
Configuration: Active config files for each component
Network Ports: ZMQ socket status and connection counts
IPSet Statistics: Blacklist entries and memory usage
Recent Activity: Last 5 log lines per component
Recent Blocks: Last 5 blocked IPs
Log Commands: Quick commands to tail individual logs

Component Metrics

Sniffer Metrics

# View sniffer statistics
grep "Enhanced RingBufferConsumer Statistics" /vagrant/logs/lab/sniffer.log -A 25 | tail -25

Output:

═══════════════════════════════════════════════════════════════════
Enhanced RingBufferConsumer Statistics:
═══════════════════════════════════════════════════════════════════
Packets processed:         152,847
Packets sent via ZMQ:      15,284 (batches)
Batch size:                10 packets/batch
Processing rate:           1,528 packets/sec
Total runtime:             100.2 seconds

Feature Groups Extracted:
  - RandomForest:          152,847 packets (52 features)
  - Ransomware:            152,847 packets (15 features)
  - Internal Anomaly:      152,847 packets (8 features)
  - DDoS Detection:        152,847 packets (8 features)

eBPF Statistics:
  - Kernel packets:        152,847
  - Dropped packets:       0
  - Ring buffer full:      0
  - Poll timeouts:         12

ZMQ Statistics:
  - Messages sent:         15,284
  - Send failures:         0
  - Average batch size:    10.0 packets
  - Compression ratio:     4.2x (LZ4)

ML Detector Metrics

# View detector statistics
grep "Stats:" /vagrant/logs/lab/detector.log | tail -10

Output:

Stats: Processed=15284 batches | DDoS=1247 | Ransomware=42 | Traffic=14891 | Internal=104 | Threats=1393
Stats: Avg latency=0.8μs | Throughput=1528 pkt/s | Memory=256MB RSS | CPU=8.5%

ML Detector Embedded Models:
  Level 1 (DDoS):          Threshold=0.85 | Detections=1247 (8.2%)
  Level 2 (Ransomware):    Threshold=0.90 | Detections=42 (0.3%)
  Level 3 (Traffic):       Threshold=0.80 | Detections=14891 (97.4%)
  Level 4 (Internal):      Threshold=0.85 | Detections=104 (0.7%)

Crypto Pipeline:
  - Encryption:            15284 messages
  - Failures:              0 (0.0%)
  - Avg encrypt time:      12.4μs

Compression Pipeline:
  - Compression:           15284 messages
  - Avg ratio:             4.2x
  - Avg compress time:     8.7μs

Firewall ACL Agent Metrics

# View firewall statistics
grep -E "Stats|Metrics" /vagrant/logs/lab/firewall-agent.log | tail -20

# Or view JSON metrics export
cat /vagrant/logs/lab/firewall-metrics.json | jq

Output:

{
  "timestamp": "2026-02-08T14:32:45Z",
  "component": "firewall-acl-agent",
  "version": "1.2.1",
  "uptime_seconds": 8100,
  "zmq": {
    "messages_received": 1393,
    "messages_processed": 1393,
    "messages_failed": 0,
    "avg_message_size_bytes": 1247,
    "connection_status": "connected"
  },
  "crypto": {
    "decryption_operations": 1393,
    "decryption_failures": 0,
    "avg_decryption_time_us": 15.2,
    "total_decryption_time_ms": 21187
  },
  "compression": {
    "decompression_operations": 1393,
    "decompression_failures": 0,
    "avg_decompression_time_us": 11.8,
    "total_decompressed_bytes": 1737851
  },
  "ipset": {
    "add_operations": 1393,
    "add_successes": 1393,
    "add_failures": 0,
    "current_entries": 1042,
    "max_capacity": 1000,
    "memory_usage_bytes": 52416
  },
  "performance": {
    "avg_processing_time_ms": 2.4,
    "max_queue_depth": 12,
    "cpu_percent": 5.2,
    "memory_rss_mb": 127
  }
}

IPSet Monitoring

View IPSet Contents

# List all IPSets
sudo ipset list -n

# View specific IPSet details
sudo ipset list ml_defender_blacklist_test

# View IPSet with statistics
sudo ipset list ml_defender_blacklist_test -t

# Count entries
sudo ipset list ml_defender_blacklist_test | grep -c "^[0-9]"

# View recent entries (last 10)
sudo ipset list ml_defender_blacklist_test | grep "^[0-9]" | tail -10

Monitor IPSet Changes

# Watch IPSet in real-time (updates every 1 second)
watch -n 1 'sudo ipset list ml_defender_blacklist_test | head -20'

# Monitor entry count
watch -n 1 'echo "Blacklist entries: $(sudo ipset list ml_defender_blacklist_test | grep -c "^[0-9]")"'

IPSet Statistics

# Get IPSet statistics
sudo ipset list ml_defender_blacklist_test -t

# Output:
Name: ml_defender_blacklist_test
Type: hash:ip
Revision: 4
Header: family inet hashsize 1024 maxelem 1000 timeout 3600
Size in memory: 52416
References: 1
Number of entries: 1042
Members:
192.168.1.100 timeout 3456
10.0.0.50 timeout 3289
...

ZeroMQ Traffic Monitoring

Port Status

# Check ZMQ ports are listening
ss -tlnp | grep -E "(5571|5572|2379|2380)"

# Check established connections
ss -tnp | grep -E "(5571|5572)" | grep ESTAB

# Monitor connection count
watch -n 1 'ss -tnp | grep 5572 | grep ESTAB | wc -l'

Message Flow

# Monitor sniffer sending to detector
grep "Sent batch" /vagrant/logs/lab/sniffer.log | tail -10

# Monitor detector receiving from sniffer
grep "Received batch" /vagrant/logs/lab/detector.log | tail -10

# Monitor detector sending to firewall
grep "Published threat" /vagrant/logs/lab/detector.log | tail -10

# Monitor firewall receiving from detector
grep "Received message" /vagrant/logs/lab/firewall-agent.log | tail -10

ZMQ Performance

# Calculate message rate (sniffer → detector)
grep "Sent batch" /vagrant/logs/lab/sniffer.log | tail -1000 | \
  awk '{print $1, $2}' | uniq -c | awk '{sum+=$1} END {print sum/NR " batches/sec"}'

# Calculate throughput (detector → firewall)
grep "Published threat" /vagrant/logs/lab/detector.log | tail -1000 | \
  awk '{print $1, $2}' | uniq -c | awk '{sum+=$1} END {print sum/NR " threats/sec"}'

Real Monitoring Scripts

ML Defender includes production-ready monitoring scripts:

monitor_lab.sh

Comprehensive dashboard (covered above):

source/scripts/monitor_lab.sh

#!/bin/bash
# ML Defender - Lab Monitoring Script (Enhanced v2.3)
# Shows: CPU, RAM, ZMQ ports, IPSet stats, config files, uptime, logs

PROJECT_ROOT="/vagrant"
LOG_DIR="$PROJECT_ROOT/logs/lab"

# Main monitoring loop
while true; do
  TIMESTAMP=$(date '+%Y-%m-%d %H:%M:%S')
  SYSTEM_UPTIME=$(uptime -p | sed 's/up //')

  # Display header
  clear
  echo "╔════════════════════════════════════════════════════════════╗"
  echo "║  ML Defender Lab - Live Monitoring (Enhanced v2.3)         ║"
  echo "║  $TIMESTAMP                                ║"
  echo "║  System Uptime: $SYSTEM_UPTIME                            "
  echo "╚════════════════════════════════════════════════════════════╝"

  # Show system stats, component status, ports, IPSet, logs
  # ... (full script in source)

  sleep 3
done

monitor_stability.sh

Long-term stability monitoring:

# Monitor stability over extended period
bash scripts/monitor_stability.sh

# Output:
Timestamp,Sniffer_PID,Detector_PID,Firewall_PID,Sniffer_CPU,Detector_CPU,Firewall_CPU,Sniffer_MEM,Detector_MEM,Firewall_MEM
2026-02-08 14:00:00,12347,12346,12345,12.1,8.5,5.2,189,256,127
2026-02-08 14:05:00,12347,12346,12345,11.8,8.7,5.1,189,257,127
...

monitor_stress.sh

Stress test monitoring:

# Monitor during stress test
bash monitor_stress.sh <sniffer_pid> <detector_pid> <output_dir> <interval>

# Output files:
# - cpu.csv: CPU usage over time
# - memory.csv: Memory usage over time
# - performance.csv: Processing rates

Log Analysis

Search for Errors

# Search all logs for errors
grep -i "error" /vagrant/logs/lab/*.log

# Search for warnings
grep -i "warning" /vagrant/logs/lab/*.log

# Search for failures
grep -i "failed" /vagrant/logs/lab/*.log

Analyze Detection Patterns

# Count detections by type
grep "Detection:" /vagrant/logs/lab/detector.log | \
  awk '{print $4}' | sort | uniq -c

# Output:
   1247 DDoS
     42 Ransomware
  14891 Traffic
    104 Internal

# View high-confidence threats
grep "Detection:" /vagrant/logs/lab/detector.log | \
  awk '$6 > 0.95' | tail -20

Analyze Blocked IPs

# Extract all blocked IPs from firewall log
grep "Blocked IP" /vagrant/logs/lab/firewall-agent.log | \
  awk '{print $5}' | sort | uniq -c | sort -rn

# Output:
     42 192.168.1.100
     28 10.0.0.50
     15 172.16.1.200
      8 203.0.113.45
      3 198.51.100.123

# Check if specific IP was blocked
grep "192.168.1.100" /vagrant/logs/lab/firewall-agent.log

Performance Analysis

# Extract processing times
grep "Processing time" /vagrant/logs/lab/detector.log | \
  awk '{print $NF}' | sed 's/ms//' | \
  awk '{sum+=$1; count+=1} END {print "Avg: " sum/count " ms"}'

# Extract throughput over time
grep "Throughput" /vagrant/logs/lab/detector.log | \
  awk '{print $1, $2, $NF}'

Dashboard and Alerting (Roadmap)

These features are planned for future releases.

Prometheus Integration (Planned)

# Future: Prometheus metrics exporter
metrics:
  prometheus:
    enabled: true
    port: 9090
    path: /metrics
    interval_seconds: 15

Grafana Dashboards (Planned)

Real-time component health
Detection rate trends
IPSet capacity utilization
ZMQ message flow
CPU/Memory usage graphs
Threat heatmaps

Alerting Rules (Planned)

# Future: Alerting configuration
alerts:
  - name: high_cpu
    condition: cpu_percent > 80
    duration: 5m
    action: email

  - name: ipset_capacity
    condition: ipset_entries > 900
    threshold: 90%
    action: slack

  - name: component_down
    condition: process_status == stopped
    duration: 1m
    action: pagerduty

Quick Reference

Monitoring Commands

# Live dashboard
bash scripts/monitor_lab.sh

# Component status
pgrep -a firewall-acl-agent
pgrep -a ml-detector
pgrep -a sniffer

# IPSet monitoring
sudo ipset list ml_defender_blacklist_test
watch -n 1 'sudo ipset list ml_defender_blacklist_test | head -20'

# Log monitoring
tail -f /vagrant/logs/lab/firewall-agent.log
tail -f /vagrant/logs/lab/detector.log
tail -f /vagrant/logs/lab/sniffer.log

# Port monitoring
ss -tlnp | grep -E "(5571|5572)"

# Performance monitoring
top -b -n 1 | grep -E "(sniffer|ml-detector|firewall)"

Vagrant Aliases

# Component logs
logs-firewall    # tail -f firewall.log
logs-detector    # tail -f detector.log
logs-sniffer     # tail -f sniffer.log
logs-lab         # Live monitoring dashboard

# Component status
status-lab       # pgrep all components

Next Steps

Performance Tuning

Optimize component performance

Troubleshooting

Diagnose and fix issues

Configuration

Configure monitoring settings

Architecture

Understand component interactions

​Log Locations

​Vagrant/Development

​Docker Compose

​Debian Package

​Monitoring Dashboard

​Live Monitoring Script

​Dashboard Features

​Component Metrics

​Sniffer Metrics

​ML Detector Metrics

​Firewall ACL Agent Metrics

​IPSet Monitoring

​View IPSet Contents

​Monitor IPSet Changes

​IPSet Statistics

​ZeroMQ Traffic Monitoring

​Port Status

​Message Flow

​ZMQ Performance

​Real Monitoring Scripts

​monitor_lab.sh

​monitor_stability.sh

​monitor_stress.sh

​Log Analysis

​Search for Errors

​Analyze Detection Patterns

​Analyze Blocked IPs

​Performance Analysis

​Dashboard and Alerting (Roadmap)

​Prometheus Integration (Planned)

​Grafana Dashboards (Planned)

​Alerting Rules (Planned)

​Quick Reference

​Monitoring Commands

​Vagrant Aliases

​Next Steps

Performance Tuning

Troubleshooting

Configuration

Architecture

Log Locations

Vagrant/Development

Docker Compose

Debian Package

Monitoring Dashboard

Live Monitoring Script

Dashboard Features

Component Metrics

Sniffer Metrics

ML Detector Metrics

Firewall ACL Agent Metrics

IPSet Monitoring

View IPSet Contents

Monitor IPSet Changes

IPSet Statistics

ZeroMQ Traffic Monitoring

Port Status

Message Flow

ZMQ Performance

Real Monitoring Scripts

monitor_lab.sh

monitor_stability.sh

monitor_stress.sh

Log Analysis

Search for Errors

Analyze Detection Patterns

Analyze Blocked IPs

Performance Analysis

Dashboard and Alerting (Roadmap)

Prometheus Integration (Planned)

Grafana Dashboards (Planned)

Alerting Rules (Planned)

Quick Reference

Monitoring Commands

Vagrant Aliases

Next Steps