syswatch: one TUI to replace htop, btop, and nvitop

Most system monitoring tools make you choose. htop for processes. btop for an overview. nvitop if you have a GPU. Swap between panes, lose context, forget what the CPU was doing three seconds ago. syswatch treats that as a solvable problem.

syswatch is a single-host system diagnostics TUI written in Rust. It monitors CPU, memory, disks, filesystems, processes, GPU, power, and network — all in one screen — and overlays graph timelines for CPU/GPU utilization so you see historical context, not just the current snapshot. 1

MIT license. 611 stars on GitHub. Version v0.7.2, released 2026-06-10 — the latest push landed yesterday. 1

github.com · GitHub 저장소

odot2/syswatch

https://github.com/odot2/syswatch

콘텐츠 카드를 불러오는 중…

What it actually monitors

The coverage is broad enough to replace multiple separate tools:

CPU: per-core utilization graph timeline plus aggregate usage — you see which cores are hot and for how long
GPU: timeline graph alongside CPU so you can correlate workload spikes
Memory: live usage with historical view
Disks and filesystems: mounted volumes, I/O activity
Processes: filterable, sortable list with pause and tick-rate control
Network: per-interface stats
Power and temperature: graphs available without root — no sudo required 1

The insights panel is worth calling out separately: it's a configurable event monitor that triggers alerts when system metrics match filters or rules you define. If you want to catch when a process pegs a core above 90% for more than ten seconds, that's the surface to set it up. 1

syswatch running in a terminal — system diagnostics dashboard showing CPU graph timeline, memory, disk, and process list — Terminal Trove preview of syswatch 1

There's also a session recorder: you can record your system monitor and replay sessions. Useful when you're trying to reproduce a spike that happened while you were away, or to hand off a problem trace to a colleague without them needing the live system.

Install

cargo install syswatch

Requires a Rust toolchain. If you don't have one: curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh, then retry the install. 1

No Homebrew formula, no apt package, no prebuilt binaries as of v0.7.2. The cargo install path is the only supported route. For engineers with Rust already installed, that's a two-second command; for everyone else, the toolchain install adds a few minutes but stays entirely local.

A scenario: diagnosing a machine after a workload spike

Your CI runner finished a build that took three times longer than expected. You ssh in and want to understand what happened — but the workload is over. Open syswatch and look at the CPU timeline. If the graph shows a single core pegged at 100% for the duration while others idled, that's a single-threaded bottleneck. If all cores were maxed out simultaneously, the job was parallelized but the machine was just saturated. The temperature and power graphs give you secondary confirmation: did the system throttle partway through?

This is the kind of retrospective trace htop can't give you — it only shows the present. syswatch's timeline graphs make the last N seconds visible even after the fact.

For ongoing work: leave syswatch open in a tmux pane while you build, run tests, or do anything compute-heavy. The per-core graph makes it immediately obvious whether your tool is using parallelism effectively.

Momentum signal

611 stars, released 2026-06-10 — v0.7.2 dropped yesterday. 1 The tool appeared on Terminal Trove, the curated CLI/TUI directory, placing it in front of the community of engineers who specifically track this category. No Hacker News thread or Reddit post is indexed yet — the tool is new enough that community discussion is just starting.

GitHub Stars

611

Version

v0.7.2

Language

Rust

License

MIT

Released

2026-06-10

통계 카드를 불러오는 중…

How it compares to btop / htop / nvitop

Terminal Trove's own description positions syswatch as a tool for users who need something "beyond or combining the capabilities of btop, htop and nvitop." 1 That framing is accurate: htop is process-centric with no GPU and no power monitoring; btop adds GPU on some systems but has no configurable event alerting; nvitop is GPU-focused and doesn't replace a full system dashboard.

syswatch attempts the consolidation. Whether the implementation matches the ambition is something you'll verify yourself once it's installed — 611 stars on day-one puts it in a credible starting position, not a proven one.

Caveats

Rust toolchain required, no prebuilts. This is the real barrier. Engineers who don't have cargo on a given machine need a toolchain install before they see anything.
v0.7.2, released yesterday. A young v0.x release. Bug surface is unknown; the feature set looks complete, but production hardening is unproven at this star count.
No community discussion yet. Zero HN/Reddit/Lobsters threads. You're going in without practitioner testimonials about edge cases or failure modes.
Single-host only. No network aggregation — this is for monitoring the machine you're on, not a fleet.

Install: cargo install syswatch (Rust toolchain required)

Cover image: AI-generated illustration

참고 출처

1syswatch — Terminal Trove