Release Notes

serverpeek 1.3.0 Beta 1 - 18 Jun 2026

Faster, steadier updates on busy servers — snapshot collectors (CPU, memory, disk, processes, Docker, network) now run concurrently, so a cycle costs as much as the slowest single source rather than the sum of them all
Docker stats and per-container top calls now run in parallel instead of one-at-a-time
Update interval is now measured from the start of collection, so the period stays close to 2 seconds instead of drifting out to 2 seconds plus however long collection took

Added a Disk card to the dashboard — shows used/total, free space, and a colour-coded usage bar for the main disk
Reports the whole-disk figure via statfs on the root mount; deliberately ignores virtual filesystems, overlays, and external drives

Moved to new GitHub location: https://github.com/WaterJuice/serverpeek

Collector now runs continuously instead of sleeping when no clients are connected, so new connections immediately see a populated history graph

Initial release
Live-updating web dashboard for server monitoring via Server-Sent Events
Dark-themed, kiosk-friendly single-page UI — no scrolling, fills the viewport
Monitors CPU usage (overall and per-core), memory, swap, and system uptime
Top processes sorted by combined CPU + memory usage, grouped by parent
Friendly process names for interpreters (Python, Node, Ruby, Perl, Java)
Docker container monitoring with stats and internal process expansion
Open network port detection via lsof (excludes localhost-only bindings)
Written in Go with zero external dependencies — stdlib only
Distributed as platform-specific Python wheels via bin2whl (pip install serverpeek)
Cross-compiled for macOS (arm64/amd64), Linux (arm64/amd64), Windows (arm64/amd64)
Statically linked single binary — no Python runtime needed at execution time
Linux distro detection via /etc/os-release (e.g. "Ubuntu 24.04")
macOS version display using kern.osproductversion (e.g. "26.4")