Dashboard tab: CPU/mem/load, hwmon temps+fans, net, disks, filesystems, UPS with 10-min sparklines

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

PLAN.md

@@ -128,10 +128,13 @@ something long in one, switch away and back, refresh — it's all still there.
 - **ZFS tab** — pools, datasets (`used/avail/refer/usedbysnapshots`), snapshot list,
   scrub status, and the `freeing` counter. `grid` is the target, which implies
   multi-host.
-- **Dashboard** — CPU/mem/load/net, disks and SMART (`smartctl -j`), temps and fans
-  (`sensors -j`), UPS from `pwrstat`. Ham Lab collects on a timer into a SQLite
-  ring (e.g. 10 s resolution, 7 days) and draws sparklines. This is where the
-  "collects its own data" requirement lands; no Prometheus dependency.
+- **Dashboard** — **first cut shipped 2026-09-17** (`internal/metrics`, `Dashboard.tsx`):
+  CPU (total + per core), memory/swap, load, uptime, hwmon temps and fans read straight
+  from `/sys` (no lm-sensors), per-interface network rates, per-disk read/write/busy
+  from `/proc/diskstats`, filesystem usage, UPS via `sudo -n pwrstat -status`. Sampled
+  every 2 s into an in-memory ring of 10 min; the UI seeds from
+  `GET /api/metrics/history` then polls `GET /api/metrics`. Still to do: persistent
+  history (SQLite ring, e.g. 10 s × 7 days), SMART (`smartctl -j`), longer time ranges.
 - **Multi-host** — same binary as a headless *agent* on `grid` / Mac mini /
   `wyse-5070-1`, `core`'s instance is the UI and fans out over mTLS. The rail grows a
   host level at the top, exactly like Proxmox's datacenter → node tree. Design the