fix(diskio): alert on the bitrate, and decorate the field the WebUI reads

[ntdatt812]

Two things, one line apart, and the network plugin next door already does both correctly.

The alert measured a lifetime counter against a bitrate threshold

alert_rx = self.get_alert(i['read_bytes'], header='rx', action_key=disk_real_name)

read_bytes carries rate: True in fields_description, which is what produces read_bytes_rate_per_sec — read_bytes itself is psutil's counter since boot. It only ever grows, so on any machine with real uptime a configured ..._rx_critical was crossed within seconds of the first read and stayed crossed for the life of the process, regardless of what the disk was doing. An operator who set thresholds got a permanently red column; one who set none saw nothing, which is why this could sit unnoticed.

The network plugin is the shape this should have had:

bps_rx = int(i['bytes_recv_rate_per_sec'] * 8)
alert_rx = self.get_alert(bps_rx, header='rx', action_key=if_real_name)

The decoration went to a key one interface reads and the other does not

update_views set it only on read_bytes / write_bytes:

• the TUI prints read_bytes_rate_per_sec (auto_unit(i.get('read_bytes_rate_per_sec'))) but reads read_bytes' decoration — a colour computed from a different quantity than the number beside it;
• the WebUI asks for getDecoration(disk.name, 'read_bytes_rate_per_sec'), which nothing ever set, so it rendered no colour at all.

Network sets both keys for exactly this reason:

self.views[...]['bytes_recv']['decoration'] = alert_rx
self.views[...]['bytes_recv_rate_per_sec']['decoration'] = alert_rx

Diskio now does the same, guarded: the _rate_per_sec view only exists once there is a timespan to divide by, and setting it unconditionally raises KeyError on the first sample — the repo's existing TestDiskioPluginViews cases catch that, which is how I found it.

Tests

tests/test_plugin_diskio.py, new TestDiskioBitrateAlert, four cases over injected stats carrying a large lifetime counter and a separate rate:

• quiet disk (rate 10) with 5 GB of history → OK — this is the case the old code got wrong
• busy disk (rate 95) → CRITICAL
• read_bytes_rate_per_sec decorated CRITICAL while write_bytes_rate_per_sec is OK — the WebUI's key, and asymmetric so a blanket assignment cannot pass it
• first sample with no rate field at all → no KeyError

Thresholds in the fixture are named diskio_sda_rx_*: get_stat_name() composes plugin + action_key + header, so a per-disk threshold carries the disk name, exactly as the network plugin's docstring documents with network_wlan0_rx_careful.

Mutation-checked, one per defect:

mutation	result
alert on i['read_bytes'] again	2 fail — the quiet-disk case and the WebUI-key case
stop setting *_rate_per_sec	1 fail — only the WebUI-key case

python -m pytest tests/test_plugin_diskio.py -q
41 passed

(37 before this PR, plus the 4 added.) Python only, so no WebUI rebuild — the bundle already reads whatever the plugin publishes.

[nicolargo]

Thanks for the PR @ntdatt812

Merged into develop.