Host metrics are the resource level measurements collected from inside a machine: CPU time per state, memory used and available, filesystem capacity per mount point, disk read and write bytes per device, network bytes in and out per interface, plus host metadata such as hostname, OS, kernel, architecture, CPU model, and boot time.
Host metrics are white box signals. They come from inside the system and explain why something is slow, in contrast to black box checks that probe from the outside and tell you whether users are affected. It is worth being precise about the boundary. A host metrics agent describes the machine as a whole, not individual processes, containers, or pods, and it carries no request traces, no logs, and no application spans. That boundary keeps the data cheap, low cardinality, and consistent across every host in a fleet, which is what makes it dependable for alerting and capacity planning.
The canonical naming for these measurements comes from the OpenTelemetry hostmetrics receiver, so the same metric name means the same thing from one vendor to the next. Hyperping's agent follows those semantic conventions and ships the full set every 30 seconds, which means historical dashboards stay meaningful even if the collector changes later.
