# Glances API

> Glances CLI

- Skill: `phattbeats/glances-api` (Agent Skill, multi-file: 8 files)
- Install (CLI): `npx skillmds@latest add phattbeats/glances-api`
- Raw SKILL.md: https://api.skillmd.com/api/skills/phattbeats/glances-api/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Integrations & APIs
- Author: phattbeats (https://skillmd.com/u/phattbeats)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/phattbeats/glances-api

---


# Glances CLI

Python CLI for the Glances monitoring API on PHATT-RAID (`10.0.0.100:61208`). No auth, instant responses, auto-detecting output.

**Why:** The old approach used 5 separate bash/node scripts piped together. One CLI, same data, cleaner.

## Quick Reference

```bash
glances.py overview             # Health check — alerts if anything's wrong
glances.py cpu                  # CPU usage + iowait
glances.py mem                  # RAM usage
glances.py temp                 # CPU temperature sensors
glances.py load                 # System load (1/5/15 min)
glances.py containers [--filter NAME] [--limit N]  # Docker containers by CPU
glances.py fs                   # Disk usage by mount
glances.py diskio               # Active disk I/O
glances.py network              # Active network interfaces
glances.py system               # Hostname, OS, uptime
```

## Setup

```bash
export GLANCES_BASE_URL="http://10.0.0.100:61208/api/4"
```

Default hardcoded — works without env vars.

## Output

Auto-detects: JSON when piped, human-readable in terminal. Force JSON with `--json`.

**JSON envelope:**
```json
{"ok": true, "command": "cpu", "result": {...}}
```

## Thresholds

| Metric | 🟡 Warning | 🔴 Critical |
|--------|-----------|------------|
| CPU % | 85 | 95 |
| RAM % | 80 | 90 |
| Temp °C | 80 | 90 |
| FS % | 85 | 95 |
| Load 1m | 10 | 12 |

Status icons in human output: 🟢 OK, 🟡 Warning, 🔴 Critical.

## Workflows

### Quick health check
```bash
glances.py overview
```
Returns CPU, memory, load, temp, container status, and any alerts.

### What's eating CPU?
```bash
glances.py containers --limit 5
```
Containers sorted by CPU usage, highest first.

### Is it hot in here?
```bash
glances.py temp
```

### Disk space check
```bash
glances.py fs
```
All mounts sorted by usage %.

### What's hitting the disks?
```bash
glances.py diskio
```

## Critical Container Monitoring

`overview` and `containers` commands check for these critical services:
- OpenClaw, swag, nextcloud, browserless, PlexMediaServer

If any are missing from the running list, an alert is raised.

## Notes

- 5 second timeout per API call
- Container CPU % can exceed 100% (multi-core)
- `diskio` and `network` only show active interfaces (with traffic)
- Temp uses AMD Tctl sensor (the one that matters for throttling)
- The old scripts (`glances_fetch.sh`, `glances_parse.mjs`, etc.) still work but this CLI replaces them