If the frequency update the prometheus.yml is faster than scrape, it will never to scrape data

[shanquanqiang]

If the frequency update the prometheus.yml is faster than scrape, it will never to scrape data.

it will return in case <-sl.scrapeCtx.Done():

func (sl *scrapeLoop) run(interval, timeout time.Duration, errc chan<- error) {
select {
case <-time.After(sl.scraper.offset(interval)):
// Continue after a scraping offset.
case <-sl.scrapeCtx.Done():
close(sl.stopped)
return
}

[fabxc]

Is this something you are interpreting from the code or something you observed in a running instance. The called offest() method does not make us wait for interval here. Instead it divides time into consistent slots and returns the time until the next one. Thus reloading frequently does not always reset the wait timer to interval and should not result in scrapes never happening.

Note that you should generally not have a need to reload Prometheus frequently regardless.

[brian-brazil]

If this was a problem, it should no longer be the case with the latest versions.

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.