Metrics and Surfaces

UnyPort is organized around a small number of operational surfaces backed by live data. The backend samples the system every 2 seconds and pushes snapshots over SSE, while the frontend renders those snapshots as focused operator pages.

Live data model

The live pipeline has a few important characteristics:

One SSE endpoint: /sse/system
One in-memory ring of 60 snapshots
Roughly 2 minutes of rolling live context
Chart scales computed server-side from the last 15 snapshots

This keeps the frontend light and avoids duplicating telemetry math in the browser.

Dashboard

The dashboard is the first reading surface:

Hostname and host role
Uptime
CPU and memory summaries
Quick links toward network, storage and security
Yearly restart heatmap derived from startup-history.jsonl or unyport.log

Hypervisor

The hypervisor page combines system identity and platform context:

Alpine and kernel version reading
Host role and runtime
BIOS and board data when available
Version comparison against TRINITY boot tags fetched from GitHub
Xen hypervisor information on Dom0
Active Xen domains on Dom0

Resources

The resources page is the broad inspection surface:

Load average
Temperatures
Top processes
Package inventory summary
Kernel modules summary
OpenRC service list and status
Allowed log-file browsing and tailing

Network and storage

The Network page shows:

Main interface
IP address
RX and TX rates
Byte totals
A network map built from host interfaces

The Storage page shows:

Mounted disks
Used and free space
Filesystem type
LBU persistence state on Alpine hosts

Security

The security page is a dedicated operational surface, not just a badge summary. It aggregates:

Kernel hardening checks
Users file permissions
OpenRC service state
Watched process presence
Listening TCP ports

That makes UnyPort more than a resource viewer. It is also a compact security status reader.