Allow scrape to be forced

Prometheus server
1

During incidents, it is sometimes good to know what is the current state of the metrics. With periodic scraping we are at the mercy of the next scrape interval to trigger. Especially for gauges which are lossy

Does prometheus allow you to force a scrape on a target?

2

No. For most metrics the scrape interval is pretty low anyway (e.g. 30s or 10s).

3

What is the work-a-round? My scrape interval is 30m since it makes db queries (via sql_exporter) which can take a while to run (requires a 25m timeout). I would like to have something setup so that when prometheus reloads its config with new sql_exporter queries, it doesn’t make me wait half an hour to see those new metrics.

prometheus.yml

global:
  scrape_interval: 30m
  scrape_timeout: 25m
...

sql_exporter.yml

global:
  min_interval: 60m
  scrape_timeout: 25m
...