Description
The host_resource probe collects host-level resource metrics: CPU usage, memory usage, disk utilization, disk I/O, and network traffic. This is a pure collector — it emits gauge/counter metrics only, no check result.
Requires the HostAdapter platform adapter — supported on Linux and Windows only. If the adapter is unavailable on the current platform, the probe is skipped at construction.
Config fields
| Field | Type | Required | Default | Description |
|---|
type | string | Yes | — | Must be "host_resource" |
collectCpu | bool | No | false | Enable CPU usage percent collection |
collectMemory | bool | No | false | Enable memory usage percent collection |
collectDisk | bool | No | false | Enable disk usage percent collection per mount point |
collectDiskIo | bool | No | false | Enable disk read/write bytes collection per device |
collectNetwork | bool | No | false | Enable network rx/tx bytes, errors, drops collection per interface |
diskMounts | string[] | No | all mounts | Mount points to monitor. Empty = all |
diskDevices | string[] | No | all devices | Disk devices for I/O monitoring. Empty = all real block devices (skips loop/ram) |
networkInterfaces | string[] | No | all interfaces | Network interfaces to monitor. Empty = all |
At least one collect* flag must be enabled, otherwise the probe fails at construction.
Metrics
| Metric | Type | Unit | Labels | Condition |
|---|
vms.host.cpu.usage_percent | gauge | percent | — | collectCpu = true |
vms.host.memory.usage_percent | gauge | percent | — | collectMemory = true |
vms.host.disk.usage_percent | gauge | percent | mount | collectDisk = true |
vms.host.disk.read_bytes | counter | bytes | device | collectDiskIo = true |
vms.host.disk.write_bytes | counter | bytes | device | collectDiskIo = true |
vms.host.network.rx_bytes | counter | bytes | iface | collectNetwork = true |
vms.host.network.tx_bytes | counter | bytes | iface | collectNetwork = true |
vms.host.network.rx_errors | counter | — | iface | collectNetwork = true |
vms.host.network.rx_drops | counter | — | iface | collectNetwork = true |
vms.host.network.tx_errors | counter | — | iface | collectNetwork = true |
vms.host.network.tx_drops | counter | — | iface | collectNetwork = true |
vms.host.uptime_seconds | gauge | seconds | — | Always collected when supported |
Example config
{
"type": "host_resource",
"collectCpu": true,
"collectMemory": true,
"collectDisk": true,
"collectDiskIo": true,
"collectNetwork": true,
"diskMounts": ["/", "/data"],
"diskDevices": ["sda", "nvme0n1"],
"networkInterfaces": ["eth0", "bond0"]
}
{
"type": "host_resource",
"collectCpu": true,
"collectMemory": true
}
Notes
- If one metric family fails (e.g. nonexistent disk mount), the probe continues collecting other families and reports
degraded status rather than failing entirely.
- Counter metrics (disk I/O, network) are cumulative since boot — downstream computes rates by differencing consecutive samples.
