Skip to content

Instantly share code, notes, and snippets.

@psiborg
Created September 20, 2026 06:19
Show Gist options
  • Select an option

  • Save psiborg/f58e1f736f30c99351e0bd842eab264c to your computer and use it in GitHub Desktop.

Select an option

Save psiborg/f58e1f736f30c99351e0bd842eab264c to your computer and use it in GitHub Desktop.
Top Tools

Top Tools

General-purpose system monitors

Tool Written in Interface Mouse Graphs Standout features Best for
top C (procps-ng) Plain text No No Installed everywhere, scriptable batch mode (top -b) Servers, minimal systems, quick checks
htop C Colored ncurses Yes Simple meters Tree view, scrolling, kill/renice from UI, customizable meters Everyday interactive use
btop (btop++) C++ Rich TUI Yes Yes Themes, CPU/mem/disk/net/GPU panels, very polished Best-looking all-in-one dashboard
bpytop Python Rich TUI Yes Yes Predecessor to btop, same look Legacy (superseded by btop)
bashtop Bash Rich TUI Yes Yes The original of the btop line Legacy (superseded by btop)
atop C Text No No Logs history to disk so you can replay past load; per-process disk/net Post-mortem analysis of past incidents
glances Python Colored TUI + web No Minimal Web UI, REST API, exports to InfluxDB/Prometheus, container stats Remote/headless monitoring
nmon C ncurses No Bar charts Toggle panels by key, CSV capture for later analysis Performance capture on IBM/enterprise boxes
bottom (btm) Rust Rich TUI Yes Yes Fast, zoomable graphs, process search/filter, configurable layout Modern lightweight alternative to btop
zenith Rust TUI Yes Yes Zoomable history charts, GPU support Charts with scrollback history
below Rust (Meta) TUI No No cgroup-aware, records and replays history Containerized/cgroup-heavy hosts
gtop Node.js TUI No Yes Dashboard-style charts Fun/visual, not production

Specialized "top" tools

Tool Focus What it shows
iotop Disk I/O Per-process read/write rates (needs root)
iftop Network Bandwidth per connection/host pair
nethogs Network Bandwidth per process
nvtop GPU NVIDIA/AMD/Intel GPU usage, memory, per-process
intel_gpu_top / radeontop GPU Vendor-specific GPU engine usage
powertop Power Wakeups and power drain per process/device, tuning suggestions
ctop Containers Docker container CPU/mem/net/IO
virt-top VMs libvirt/KVM guest resource usage
kmon Kernel Loaded kernel modules and kernel activity
s-tui CPU stress/thermals Frequency, temperature, power with built-in stress test
mytop / pg_top Databases Active MySQL/PostgreSQL queries
apachetop Web server Live Apache/nginx log request rates

Quick picks

For a server where you can't install anything, top is always there. For daily interactive use, htop is the safe default and btop is the prettiest. If you need to know what happened at 3 a.m. last night, atop or below are the ones that record history. And when the question is specifically "who's hammering the disk/network/GPU," reach for iotop, nethogs, or nvtop respectively.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment