Observability with Prometheus and Grafana

Short-Lived Jobs and the Pull Model

Prometheus, with its default setup, is not suited for monitoring short-lived jobs. If you have jobs that run for a short period of time and then terminate, Prometheus may not be able to collect metrics from them. Why? Because Prometheus collects metrics at regular intervals (e.g.: scrape_interval: 15s), and if the job terminates before Prometheus can scrape its metrics, data will be lost. In this case, you can use the Pushgateway