Try the following Prometheus subquery <https://prometheus.io/blog/2019/01/28/subquery-support/>:
count_over_time((CMSummary)[1h:1m]) == 60 It should return only time series, which were available every minute during the last hour. On Fri, May 1, 2020 at 11:59 AM Julius Volz <[email protected]> wrote: > Btw. you could even modify this expression to check manually every e.g. 15 > minutes within the last hour, whether an IP was present at that time > increment within the last hour: > > CMSummary offset 1h > and > CMSummary offset 45m > and > CMSummary offset 30m > and > CMSummary offset 15m > and > CMSummary > > So you see, theoretically you could even check every minute or so for the > presence, but that would become a long query... > > On Fri, May 1, 2020 at 10:55 AM Julius Volz <[email protected]> wrote: > >> Do you also need to exclude IPs that are present at the beginning of the >> hour, go missing briefly in between (as in, the time series becomes fully >> absent), but are present again at the end of the hour? We don't have a way >> in PromQL to check whether a series has been absent just briefly within an >> interval, but is there most of the time. We have absent_over_time(), but >> that only checks whether a series has been fully absent over a given time >> range. >> >> But if your IPs behave in such a way that they don't appear / disappear >> that rapidly, you could check which ones were there both at the beginning >> and the end of the interval: >> >> CMSummary offset 1h and CMSummary >> >> On Thu, Apr 30, 2020 at 10:51 PM Arnav Bose <[email protected]> >> wrote: >> >>> Hi, >>> >>> I know I can check in the graph how long the metric has been available. >>> In my case I want to create a query which will list down the data for a >>> metric which has been available for an hour, excluding the ones which at >>> least went down/missing during the same period. >>> >>> Here is my metric - CMSummary{ipAddr="$$$"}. There are at least 20000 >>> different IP sources with this metric. I want to know which ones have been >>> available for the last 1 hr. >>> >>> >>> Thanks, >>> Arnav >>> >>> >>> -- >>> 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/d21d5adf-0307-4367-b050-ec81cf0ce8b2%40googlegroups.com >>> <https://groups.google.com/d/msgid/prometheus-users/d21d5adf-0307-4367-b050-ec81cf0ce8b2%40googlegroups.com?utm_medium=email&utm_source=footer> >>> . >>> >> -- > 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/CA%2BT6Yow7Ka2eXx7a-VaJxDpXC8PAmTXTU3tYwnFVMzVvTe4NKA%40mail.gmail.com > <https://groups.google.com/d/msgid/prometheus-users/CA%2BT6Yow7Ka2eXx7a-VaJxDpXC8PAmTXTU3tYwnFVMzVvTe4NKA%40mail.gmail.com?utm_medium=email&utm_source=footer> > . > -- Best Regards, Aliaksandr -- 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/CAPbKnmAfaL41D13xWu%2B%3DFf4Kgv-eV8maD3zhazEt6JNVwnx6Pg%40mail.gmail.com.
