Skip to main content

OTEL-055: hostmetrics collection_interval below 10s

Severity: warn (advisory)

Rule Details​

hostmetrics reads /proc, /sys, and platform APIs on every tick to produce CPU, memory, disk, and network metrics. At intervals below 10s the cost of the collection itself starts showing up in the host metrics you are collecting — a self-measurement feedback loop that burns CPU without giving you meaningfully better resolution.

This rule fires when a hostmetrics receiver has collection_interval < 10s.

Options​

FieldConstraint
collection_intervalShould be ≥ 10s

Examples​

Avoid
receivers:
  hostmetrics:
    collection_interval: 1s           # too aggressive
    scrapers:
      cpu:
      memory:
Prefer
receivers:
  hostmetrics:
    collection_interval: 30s
    scrapers:
      cpu:
      memory:
      disk:
      network:

When Not To Use It​

A short forensic run during incident response where you deliberately want sub-10s granularity. Revert after the investigation.

  • OTEL-054 — Prometheus scrape_interval below 10s

Version​

Available since augur v0.1.0.

Further Reading​

Resources​