Configure the OpenTelemetry Collector hostmetrics receiver to collect CPU, memory, disk, filesystem, and network metrics from the host

Steps

1. Add a hostmetrics receiver block under receivers with collection_interval (default 1m) and initial_delay (default 1s), e.g. hostmetrics: { collection_interval: 30s }.
2. Under scrapers, enable the desired categories by adding their keys (empty or with options): cpu, memory, disk, filesystem, network — plus load, paging, processes, process, system, or nfs where OS-supported.
3. For filesystem, use include_devices/exclude_devices, include_fs_types/exclude_fs_types, and include_mount_points/exclude_mount_points blocks (each with match_type: strict or regexp) to scope which filesystems are scraped, and set include_virtual_filesystems: true if needed.
4. For disk and network, use include/exclude blocks with devices or interfaces plus match_type to filter which devices/NICs are scraped.
5. Add the receiver to a metrics pipeline under service.pipelines.metrics.receivers.
6. When running inside a container, bind-mount the host filesystem (e.g. docker run -v /:/hostfs) and set root_path: /hostfs so the scrapers read host paths instead of the container's own; all hostmetrics instances on the same Collector must use the same root_path.