System Health
Purpose
System Health — a live, read-only dashboard of host resource utilization, refreshing every 2 seconds: CPU, memory, disk I/O, filesystems, network interfaces, and top processes, rolled up into a single health score.
When to use it
- To get a fast, single-number (0–100) read on whether the host is under stress
- To find out what's bottlenecking the host — CPU, memory, disk, or network — before adding more VMs to it
- To identify the specific process consuming the most CPU or memory
- To check filesystem usage or per-disk I/O latency and queue depth on the physical host
How to get there
- Route / id:
/app/system-health - Nav: Infrastructure → System Health (sidebar, command palette, or desktop nav)
Operate from the console (UX)
- Health score gauge (0–100, colored red/amber/green) with a status label and a summary panel calling out the current bottleneck, if any.
- CPU panel — overall usage, core count, 1-minute and 5-minute load averages, and a per-core usage grid.
- Memory panel — RAM and swap usage bars with used/total detail, plus total, available, and cached memory.
- Filesystems — a usage bar per mounted filesystem (mountpoint, filesystem type, used/total), shown when data is available.
- Disk I/O — per-device reads/writes completed, bytes read/written, average latency, and queue depth.
- Network — per-interface RX/TX bytes and error counts, TCP connection-state counts, and host-wide RX/TX/retransmit totals.
- Top CPU and Top Memory process tables — PID, name, CPU%, and memory (MB).
Entirely read-only monitoring — there's no create/edit/delete action here; data refreshes automatically every 2 seconds.
Empty / fail: Error banner, empty table, or failed toast — confirm you are signed in, zyvor-fabricd is healthy (/readyz), and any backend this page needs (FluxVM, storage, network) is reachable. See Admin basics.
Success: Live data loads without error; creates/updates appear in the list or detail view and any confirmation toast clears cleanly.