hardware_info event: cpu_overall_percent is a misleading field name

Currently, `blitz_api` sends the following `hardware_info` event:

```json
{
"cpu_overall_percent": 1.87,
"cpu_per_cpu_percent": [1.77, 1.91, 1.92],
"vram_total_bytes": 3844000000,
"vram_available_bytes": 2587000000,
"vram_used_bytes": 1257000000,
"vram_usage_percent": 32.7,
"temperatures_celsius": { "system_temp": 56.0, "coretemp": [] },
"boot_time_timestamp": 1650185397.5461075,
"disks": [
{
"device": "/",
"mountpoint": "/",
"filesystem_type": "ext4",
"partition_total_bytes": 1000203820544,
"partition_used_bytes": 572113763840,
"partition_free_bytes": 428090056704,
"partition_percent": 42.8
}
],
"networks": {
"internet_online": "xxx",
"tor_web_addr": "xxx",
"internet_localip": "xxx",
"internet_localiprange": "xxx"
}
}
```

`cpu_overall_percent` is not the percentage of the CPU, but the load.
`cpu_per_cpu_percent` are the load averages in the last minute, last 5min, last 15min

See also https://scoutapm.com/blog/understanding-load-averages

Suggestion: rename the fields, I'll do the correct conversion in frontend :)