Description
Prometheus metrics from short-lived or batch jobs can be accepted and held until scraped. This service is for operators monitoring tasks that cannot be scraped directly while running. It is network-facing monitoring infrastructure, so label trust, stale metrics, access control, and retention policy matter.