Thank you, Stuart. If I push the data using push gateway once in 24 hours, is querying still a problem(you mentioned it is marked as stale and most queries won't work)? If I have data and want to get the most recent metrics older than 2 mins(recommended maximum scrape) lets say within 24 hours from the current time point, what query can I use? So depending on the query time, the most recent metric could be a couple of minutes older to a few hours to max 24 hours older than current time.
On Fri, Nov 13, 2020 at 4:03 AM Stuart Clark <[email protected]> wrote: > On 13/11/2020 02:47, kiran wrote: > > Hello all > > > > I have a use case where I have a metric coming once every 24 hours and > > that time varies per team. Now for each team we want to get most > > recent value of that metric. Here issue is I don’t know the offset or > > time duration as to when was the last update of that metric. So here I > > need to use offset 24h ? > > Does using offset get matches greater than 24 hours from current time > > or latest metric within past 24 hours? From documentation not able to > > figure out the definition of offset. > > > Offset will just look at 24 hours before now (or whatever the time > specified for the query). > > As with all metrics in Prometheus you need to ensure they are > successfully scraped regularly, with the maximum recommended scrape > interval being about 2 minutes. To find the latest value is then a very > simple query. > > If you scrape less frequently you will end up with a metric which is > regularly marked as stale, and therefore most queries won't work - they > just won't find any valid data and will return nothing. > > If the source of this daily change can't be scraped directly every 2 > minutes, this could be a use case for the Push gateway or the textfile > collector of the node exporter. Your daily process would publish the > metrics to either, which are then kept to allow them to be regularly > scraped. One common pattern is to also include a metric where the value > is the timestamp the process started or finished, to allow you to > detect/alert on failures of this daily process. > > -- > You received this message because you are subscribed to the Google Groups > "Prometheus Users" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/prometheus-users/81079e62-5da4-6de8-44ae-27752cfec28b%40Jahingo.com > . > -- You received this message because you are subscribed to the Google Groups "Prometheus Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/CAOnWYZVQhUV-faCENTU7Vo9cX8C0Q2oCNv4VbopaFk3NF8WPRQ%40mail.gmail.com.
