Files

kelinfoxy ef55974b50 Wiki major update

updated with recent documentation

2026-01-21 19:18:39 -05:00

15 KiB

Raw Blame History

Glances - System Monitoring Dashboard

Overview
What is Glances?
Why Use Glances?
How It Works
Configuration in AI-Homelab
Official Resources
Educational Resources
Docker Configuration
Using Glances
Advanced Topics
Troubleshooting

Overview

Category: Infrastructure Monitoring
Docker Image: nicolargo/glances
Default Stack: infrastructure.yml
Web UI: https://glances.${DOMAIN}
Authentication: Protected by Authelia (SSO)
Purpose: Real-time system resource monitoring

What is Glances?

Glances is a cross-platform system monitoring tool that provides a comprehensive overview of your system's resources. It displays CPU, memory, disk, network, and process information in a single interface, accessible via CLI, Web UI, or API.

Key Features

Comprehensive Monitoring: CPU, RAM, disk, network, sensors, processes
Real-Time Updates: Live statistics updated every few seconds
Web Interface: Beautiful responsive dashboard
REST API: Export metrics for other tools
Docker Support: Monitor containers and host system
Alerts: Configurable thresholds for warnings
Historical Data: Short-term data retention
Extensible: Plugins for additional monitoring
Export Options: InfluxDB, Prometheus, CSV, JSON
Lightweight: Minimal resource usage
Cross-Platform: Linux, macOS, Windows

Why Use Glances?

Quick Overview: See system health at a glance
Resource Monitoring: Track CPU, RAM, disk usage
Process Management: Identify resource-heavy processes
Container Monitoring: Monitor Docker containers
Network Analysis: Track network bandwidth
Temperature Monitoring: Hardware sensor data
Disk I/O: Identify disk bottlenecks
Easy Access: Web interface from any device
No Complex Setup: Works out of the box
Free & Open Source: No licensing costs

How It Works

Host System (CPU, RAM, Disk, Network)
     ↓
Glances Container (accesses host metrics via /proc, /sys)
     ↓
Data Collection & Processing
     ↓
┌─────────────┬──────────────┬────────────┐
│  Web UI     │  REST API    │  CLI       │
│  (Port 61208)│ (JSON/Export)│  (Terminal)│
└─────────────┴──────────────┴────────────┘

Monitoring Architecture

Host Access:

/proc - Process and system info
/sys - Hardware information
/var/run/docker.sock - Docker container stats
/etc/os-release - System information

Data Flow:

Collect: Gather metrics from system files
Process: Calculate rates, averages, deltas
Store: Keep short-term history in memory
Display: Render in web UI or export
Alert: Check thresholds and warn if exceeded

Configuration in AI-Homelab

Directory Structure

/opt/stacks/infrastructure/glances/
└── config/
    └── glances.conf  # Optional configuration file

Environment Variables

# Web server mode
GLANCES_OPT=-w

# Timezone
TZ=America/New_York

# Update interval (seconds)
# GLANCES_OPT=-w -t 2

# Disable web password (use Authelia)
# GLANCES_OPT=-w --disable-webui

Official Resources

Website: https://nicolargo.github.io/glances/
GitHub: https://github.com/nicolargo/glances
Docker Hub: https://hub.docker.com/r/nicolargo/glances
Documentation: https://glances.readthedocs.io
Wiki: https://github.com/nicolargo/glances/wiki

Educational Resources

Videos

Articles & Guides

Concepts to Learn

/proc filesystem: Linux process information
CPU Load Average: 1, 5, and 15-minute averages
Memory Types: RAM, swap, cache, buffers
Disk I/O: Read/write operations per second
Network Metrics: Bandwidth, packets, errors
Process States: Running, sleeping, zombie
System Sensors: Temperature, fan speeds, voltages

Docker Configuration

Complete Service Definition

glances:
  image: nicolargo/glances:latest
  container_name: glances
  restart: unless-stopped
  pid: host  # Required for accurate process monitoring
  networks:
    - traefik-network
  volumes:
    - /var/run/docker.sock:/var/run/docker.sock:ro
    - /etc/os-release:/etc/os-release:ro
    - /opt/stacks/infrastructure/glances/config:/glances/conf:ro
  environment:
    - GLANCES_OPT=-w
    - TZ=America/New_York
  labels:
    - "traefik.enable=true"
    - "traefik.http.routers.glances.rule=Host(`glances.${DOMAIN}`)"
    - "traefik.http.routers.glances.entrypoints=websecure"
    - "traefik.http.routers.glances.tls.certresolver=letsencrypt"
    - "traefik.http.routers.glances.middlewares=authelia@docker"
    - "traefik.http.services.glances.loadbalancer.server.port=61208"

Important Mount Points

volumes:
  # Docker container monitoring
  - /var/run/docker.sock:/var/run/docker.sock:ro
  
  # System information
  - /etc/os-release:/etc/os-release:ro
  
  # Host filesystem (optional, for disk monitoring)
  # - /:/rootfs:ro
  
  # Configuration file (optional)
  - /opt/stacks/infrastructure/glances/config:/glances/conf:ro

Network Modes

For better host monitoring, use host network:

glances:
  network_mode: host
  # Then access via: http://SERVER_IP:61208
  # Or still use Traefik but with host networking

Using Glances

Dashboard Overview

Web Interface Sections:

Header:
- Hostname
- System uptime
- Linux distribution
- Current time
CPU:
- Overall CPU usage (%)
- Per-core utilization
- Load average (1, 5, 15 min)
- Context switches
Memory:
- Total RAM
- Used/Free
- Cache/Buffers
- Swap usage
Disk:
- Partition information
- Space used/free
- Mount points
- Disk I/O rates
Network:
- Interface names
- Upload/Download rates
- Total transferred
- Errors and drops
Sensors:
- CPU temperature
- Fan speeds
- Other hardware sensors
Docker:
- Container list
- Container CPU/Memory
- Container I/O
Processes:
- Top CPU/Memory processes
- Process details
- Sort options

Color Coding

Glances uses colors to indicate resource usage:

Green: OK (< 50%)
Blue: Caution (50-70%)
Magenta: Warning (70-90%)
Red: Critical (> 90%)

Sorting Options

Click on column headers to sort:

CPU: CPU usage
MEM: Memory usage
TIME: Process runtime
NAME: Process name
PID: Process ID

Keyboard Shortcuts (CLI Mode)

If accessing via terminal:

a - Sort by automatic (CPU + memory)
c - Sort by CPU
m - Sort by memory
p - Sort by process name
i - Sort by I/O rate
t - Sort by time (cumulative)
d - Show/hide disk I/O
f - Show/hide filesystem
n - Show/hide network
s - Show/hide sensors
k - Kill process
h - Show help
q - Quit

Advanced Topics

Configuration File

Create custom configuration for alerts and thresholds:

glances.conf:

[global]
refresh=2
check_update=false

[cpu]
# CPU thresholds (%)
careful=50
warning=70
critical=90

[memory]
# Memory thresholds (%)
careful=50
warning=70
critical=90

[load]
# Load average thresholds
careful=1.0
warning=2.0
critical=5.0

[diskio]
# Disk I/O hide regex
hide=loop.*,ram.*

[fs]
# Filesystem hide regex
hide=/boot.*,/snap.*

[network]
# Network interface hide regex
hide=lo,docker.*

[docker]
# Show all containers
all=true
# Max containers to display
max_name_size=20

[alert]
# Alert on high CPU
cpu_careful=50
cpu_warning=70
cpu_critical=90

Mount config:

volumes:
  - /opt/stacks/infrastructure/glances/config/glances.conf:/glances/conf/glances.conf:ro

Export to InfluxDB

Send metrics to InfluxDB for long-term storage:

glances.conf:

[influxdb]
host=influxdb
port=8086
protocol=http
user=glances
password=glances
db=glances
prefix=localhost
tags=environment:homelab

Export to Prometheus

Make metrics available for Prometheus scraping:

glances:
  environment:
    - GLANCES_OPT=-w --export prometheus
  ports:
    - "9091:9091"  # Prometheus exporter port

Prometheus config:

scrape_configs:
  - job_name: 'glances'
    static_configs:
      - targets: ['glances:9091']

REST API

Access metrics programmatically:

# Get all stats
curl http://glances:61208/api/3/all

# Get specific stat
curl http://glances:61208/api/3/cpu
curl http://glances:61208/api/3/mem
curl http://glances:61208/api/3/docker

# Get process list
curl http://glances:61208/api/3/processlist

# Full API documentation
curl http://glances:61208/docs

Alerts and Actions

Configure alerts in glances.conf:

[alert]
disable=False

[process]
# Alert if process not running
list=sshd,nginx,docker
# Alert if process running
disable_pattern=.*badprocess.*

[action]
# Execute script on alert
critical_action=script:/scripts/alert.sh

Multi-Server Monitoring

Monitor multiple servers:

On each remote server:

glances:
  image: nicolargo/glances
  environment:
    - GLANCES_OPT=-s  # Server mode
  ports:
    - "61209:61209"

On main server:

glances:
  environment:
    - GLANCES_OPT=-w --browser
    # Or use client mode to connect

Access via web UI: "Remote monitoring" section

Custom Plugins

Create custom monitoring plugins:

# /opt/stacks/infrastructure/glances/config/plugins/custom_plugin.py
from glances.plugins.glances_plugin import GlancesPlugin

class Plugin(GlancesPlugin):
    def update(self):
        # Your custom monitoring code
        stats = {}
        stats['custom_metric'] = get_custom_data()
        return stats

Troubleshooting

Glances Not Showing Host Metrics

# Verify host access
docker exec glances ls /proc
docker exec glances cat /proc/cpuinfo

# Check pid mode
docker inspect glances | grep -i pid

# Ensure proper mounts
docker inspect glances | grep -A10 Mounts

# Try host network mode
# In compose: network_mode: host

Docker Containers Not Visible

# Verify Docker socket mount
docker exec glances ls -la /var/run/docker.sock

# Check permissions
docker exec glances docker ps

# Ensure docker section enabled in config
# Or no config file hiding it

High CPU Usage from Glances

# Increase refresh interval
GLANCES_OPT=-w -t 5  # Update every 5 seconds

# Disable modules you don't need
# In glances.conf:
# disable=sensors,raid

# Check if something is hammering the API
docker logs traefik | grep glances

Temperature Sensors Not Showing

# Need access to /sys
volumes:
  - /sys:/sys:ro

# Install lm-sensors on host
sudo apt install lm-sensors
sudo sensors-detect

# Verify sensors work on host
sensors

Web Interface Not Loading

# Check if Glances is running
docker ps | grep glances

# View logs
docker logs glances

# Test direct access
curl http://SERVER_IP:61208

# Check Traefik routing
docker logs traefik | grep glances

# Verify web mode enabled
docker exec glances ps aux | grep glances

Disk Information Incomplete

# Mount host root filesystem
volumes:
  - /:/rootfs:ro

# In glances.conf:
[fs]
hide=/rootfs/boot.*,/rootfs/snap.*

Memory Information Incorrect

# Use host PID namespace
pid: host

# Check /proc/meminfo access
docker exec glances cat /proc/meminfo

# Restart container
docker restart glances

Performance Optimization

Reduce Update Frequency

environment:
  - GLANCES_OPT=-w -t 5  # Update every 5 seconds (default is 2)

Disable Unnecessary Modules

glances.conf:

[global]
# Disable modules you don't need
disable=raid,sensors,hddtemp

Limit Process List

[processlist]
# Max number of processes to display
max=50

Resource Limits

glances:
  deploy:
    resources:
      limits:
        cpus: '0.5'
        memory: 256M

Security Considerations

Protect with Authelia: Exposes sensitive system info
Read-Only Mounts: Use :ro for all mounted volumes
Limited Socket Access: Consider Docker Socket Proxy
No Public Access: Never expose without authentication
API Security: Restrict API access if enabled
Process Info: Can reveal application details
Network Monitoring: Shows internal network traffic
Regular Updates: Keep Glances container updated
Audit Logs: Monitor who accesses the interface
Minimal Permissions: Only mount what's necessary

Integration with Other Tools

Grafana Dashboard

Export to Prometheus, then visualize in Grafana:

Enable Prometheus export in Glances
Add Prometheus datasource in Grafana
Import Glances dashboard: https://grafana.com/grafana/dashboards/

Home Assistant

Monitor via REST API:

sensor:
  - platform: rest
    resource: http://glances:61208/api/3/cpu
    name: Server CPU
    value_template: '{{ value_json.total }}'
    unit_of_measurement: '%'

Uptime Kuma

Monitor Glances availability:

Type: HTTP(s)
URL: https://glances.yourdomain.com
Heartbeat: Every 60 seconds

Comparison with Alternatives

Glances vs htop/top

Glances:

Web interface
Historical data
Docker monitoring
Remote access
Exportable metrics

htop:

Terminal only
No web interface
Lower overhead
More detailed process tree

Glances vs Netdata

Glances:

Simpler setup
Lighter weight
Better for single server
Python-based

Netdata:

More detailed metrics
Better long-term storage
Complex setup
Better for multiple servers

Glances vs Prometheus + Grafana

Glances:

All-in-one solution
Easier setup
Less powerful
Short-term data

Prometheus + Grafana:

Enterprise-grade
Long-term storage
Complex setup
Powerful querying

Summary

Glances provides real-time system monitoring in a simple, accessible interface. It:

Shows comprehensive system metrics at a glance
Monitors Docker containers alongside host
Provides web UI and REST API
Uses minimal resources
Requires minimal configuration
Perfect for quick system health checks

Best For:

Homelab monitoring
Quick troubleshooting
Resource usage overview
Docker container monitoring
Single-server setups

Not Ideal For:

Long-term metric storage
Complex alerting
Multi-server enterprise monitoring
Detailed performance analysis

Remember:

Protect with Authelia
Use host PID mode for accurate monitoring
Mount Docker socket for container stats
Configure thresholds for alerts
Complement with Grafana for long-term analysis
Lightweight alternative to complex monitoring stacks

15 KiB Raw Blame History

Glances - System Monitoring Dashboard

Table of Contents

Overview

What is Glances?

Key Features

Why Use Glances?

How It Works

Monitoring Architecture

Configuration in AI-Homelab

Directory Structure

Environment Variables

Official Resources

Educational Resources

Videos

Articles & Guides

Concepts to Learn

Docker Configuration

Complete Service Definition

Important Mount Points

Network Modes

Using Glances

Dashboard Overview

Color Coding

Sorting Options

Keyboard Shortcuts (CLI Mode)

Advanced Topics

Configuration File

Export to InfluxDB

Export to Prometheus

REST API

Alerts and Actions

Multi-Server Monitoring

Custom Plugins

Troubleshooting

Glances Not Showing Host Metrics

Docker Containers Not Visible

High CPU Usage from Glances

Temperature Sensors Not Showing

Web Interface Not Loading

Disk Information Incomplete

Memory Information Incorrect

Performance Optimization

Reduce Update Frequency

Disable Unnecessary Modules

Limit Process List

Resource Limits

Security Considerations

Integration with Other Tools

Grafana Dashboard

Home Assistant

Uptime Kuma

Comparison with Alternatives

Glances vs htop/top

Glances vs Netdata

Glances vs Prometheus + Grafana

Summary

15 KiB

Raw Blame History