# perf

The figures the conky dashboard used to show, as a drawer page of cards, one
per row; the page scrolls. CPU load with a bar per core, the CPU model and
RAM; every temperature as a tile; the Arc B580's VRAM and fan (no
utilisation, xe exposes none); download and upload over the last minute, the
LAN address and the public IP; the five mounts with their usage; and
`~/.cache` with its five largest children. `Perf.qml` reads them; `Perf.js`
holds the parsing, tested by `node desktop/modules/perf/test_perf.js`.

The tile shows CPU and RAM, and a dot for the worst of these, yellow or red:

| Input | Yellow | Red |
| --- | --- | --- |
| each temperature | its warn | its crit |
| RAM | 85% | 95% |
| each disk | 90% | 95% |
| public IP | | differs from the pinned one |

CPU load is left out, since spikes would flicker it. There is no hysteresis,
so a sensor sitting on its threshold can toggle the dot. The pin beside the
public IP stores the address shown as `perf.pinnedIp` in
`$XDG_STATE_HOME/keystone/desktop.json`; the check exists because the VPN
into the LAN depends on the ISP's fixed address. The fetch is
`curl -4 https://ipinfo.io/ip` every 30 minutes from the last attempt, into
`~/.cache/udt/pubip.txt`, the file conky's sampler used. A failed fetch is
never a mismatch: the old value goes stale after 4 hours and shows as a dash.

Everything is sampled always, including per-core load, the network rate,
VRAM (a scan of every fdinfo, ~14ms every 2s) and the fan, so the page opens
on the last minute of network history instead of an empty graph. The mount list, `br0`, the GPU name and the 12 GiB VRAM
size are constants in `Perf.qml`.

The cache card runs `du -B1 -d1` once each time the page opens, 28ms over
2 GiB, instead of reading conky's 15-minute `cachesize.txt`. The largest
child is yellow, red when it holds over half the total, as conky had it.