cargo-slow 0.1.0

Cargo subcommand to diagnose a slow machine: identify disk, memory, CPU, and thermal issues
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88

# cargo-slow

[![CI](https://github.com/LucaCappelletti94/cargo-slow/actions/workflows/ci.yml/badge.svg)](https://github.com/LucaCappelletti94/cargo-slow/actions/workflows/ci.yml)
[![codecov](https://codecov.io/gh/LucaCappelletti94/cargo-slow/branch/main/graph/badge.svg)](https://codecov.io/gh/LucaCappelletti94/cargo-slow)
[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
[![MSRV](https://img.shields.io/badge/MSRV-1.88-blue.svg)](https://www.rust-lang.org)

`cargo-slow` is a Linux terminal monitor for diagnosing slow machines, shipped as the `cargo slow` subcommand. It samples system pressure, temperatures, disk health, and optional throughput benchmarks, then shows the results in a TUI and writes them to CSV.

It is meant for the practical question: is this slowdown caused by I/O, memory pressure, thermal limits, disk health, or general resource exhaustion?

## Install

```bash
cargo install cargo-slow
```

This installs the `cargo-slow` binary, which Cargo runs as the `cargo slow` subcommand. You can also invoke the binary directly as `cargo-slow`.

## Usage

```bash
# TUI dashboard
cargo slow

# Full hardware data, when SMART/IPMI access needs privileges
sudo $(which cargo-slow)

# Logging only
cargo slow --headless

# Include active disk throughput checks
cargo slow --io-bench

# Show all options
cargo slow --help
```

`q`, `Esc`, and `Ctrl+C` exit the TUI. If stdout is not a terminal, `cargo slow` falls back to headless mode.

## What It Tracks

- CPU usage, load, iowait, and CPU pressure.
- Memory, swap, dirty/writeback pages, and memory pressure.
- Disk I/O counters, queue depth, and optional read/write/hash benchmarks.
- Network counters, process counts, and file descriptor usage.
- CPU, DIMM, NVMe, SATA disk, and IPMI/BMC temperatures when available.
- SMART health, reallocated sectors, pending sectors, unsafe shutdowns, and disk pass/fail state.

Implausible sensor values are ignored rather than plotted or logged as real temperatures.

## Privileges

Most `/proc` and `/sys` metrics work as a normal user. Run with `sudo` for full SMART and IPMI/BMC coverage, especially SATA disk temperatures and disk health.

Optional tools:

| Tool | Used for |
|------|----------|
| `smartctl` from `smartmontools` | SMART health and SATA disk temperatures |
| `ipmitool` | BMC/IPMI DIMM sensors |

The UI reports missing tools, missing sensors, and permission gaps.

## Options

| Option | Default | Meaning |
|--------|---------|---------|
| `-i, --interval` | `5` | Seconds between samples |
| `-c, --csv-file` | `metrics.csv` | CSV output path |
| `-t, --test-file` | `/tmp/slowtest.bin` | File used by I/O benchmarks |
| `-f, --file-size-mb` | `256` | I/O benchmark file size |
| `--history-size` | `120` | Points kept for TUI plots |
| `--headless` | off | Disable the TUI |
| `--io-bench` | off | Enable active disk throughput benchmarks |

`--io-bench` is disabled by default because it adds disk load and may disturb an already slow system. Enable it only when you want direct throughput measurements.

## Requirements

- Linux.
- Rust 1.88 or newer.
- Kernel PSI support is optional but useful for pressure metrics.
- Hardware temperature support depends on the machine and loaded kernel drivers such as `coretemp`, `k10temp`, `zenpower`, `jc42`, and NVMe hwmon drivers.

## License

MIT