Wow thats awesome the offset query is working as expected Thank you
But to push values like pressure or temperature in our use case we need to push model accuracy - we use gauge but till the next push the value is still the previous value In this scenario the only solution is to set it back to 0 ? On Thu, Feb 17, 2022 at 1:17 AM Brian Candler <[email protected]> wrote: > That's just a matter of the PromQL query used to visualize the data. Try: > > (Q23_counter - Q23_counter offset 1m) >= 0 > > Another option is increase(Q23_counter[1m]), but since it interpolates the > value over the time range given, it won't show values of exactly 1.0 > > On Thursday, 17 February 2022 at 00:10:16 UTC [email protected] wrote: > >> Hi Brian, >> >> Thank you very much for replying >> >> I meant push because I am using pushgateway in python prometheus client >> >> I will definitely consider counter for this >> >> But Out of 3 drifts - I want a spike in bar chart whenever a drift occurs >> Like this >> >> >> Regards, >> Yegna >> >> >> On Wed, Feb 16, 2022 at 12:37 Brian Candler <[email protected]> wrote: >> >>> It sounds like you're attacking this from the wrong direction. >>> >>> What I believe you want is a counter of drift events. Each time a drift >>> event occurs, you increment the counter. 0 to 1, 1 to 2, 2 to 3 etc. Never >>> reset it to 0. This is the standard and most useful way to handle such >>> data in Prometheus. Using a proper counter also handles edge cases that >>> you might not have considered, such as two of these "drift events" >>> occurring between scrapes. >>> >>> You can the write queries (e.g. for alerts) which will tell you whether >>> the counter has increased in the last 10 minutes, or the last hour, or >>> whatever. Such things *will* naturally reset to zero. For example, here's >>> an alerting rule I use: >>> >>> expr: increase(megaraid_pd_shield_counter[72h]) > 0 >>> >>> This gives an alarm whenever the counter has increased; when it *hasn't* >>> increased for 72 hours then the alert is cleared. >>> >>> P.S. Note that you don't "push" a value to prometheus; prometheus >>> "pulls" a value by scraping the exporter. >>> >>> On Wednesday, 16 February 2022 at 15:29:50 UTC [email protected] >>> wrote: >>> >>>> Hi Brian, >>>> >>>> Good day, >>>> >>>> I have a code that monitors a drift in data >>>> >>>> If drift happens I need to push value 1 to prometheus >>>> Default 0 , issue is when I set gauge as 1, It is 1 till I set it back >>>> to 0 >>>> >>>> Is there any metric where I can push the value 1 and it resets to 0 >>>> automatically ? >>>> >>>> Regards, >>>> Yegna >>>> >>>> >>>> >>>> On Fri, Jan 21, 2022 at 04:35 Brian Candler <[email protected]> wrote: >>>> >>>>> > I have considered the grafana data source option but I found a >>>>> plugin for mongoDb which is available only in enterprise grafana >>>>> > >>>>> > So do you suggest me to create a data source for mongoDB in grafana ? >>>>> >>>>> That's a grafana question, not a prometheus question; it would be >>>>> better asked in the grafana community <https://community.grafana.com/> >>>>> . >>>>> >>>>> > And regarding 2 and 3 - since we use python - Is prometheus-python >>>>> client API useful to directly write to prometheus ? (remote write >>>>> protocol) >>>>> >>>>> If you mean https://github.com/prometheus/client_python, then as far >>>>> as I know it's only an exporter, not a remote write client. Remote write >>>>> was option 3. You could send data to vector.dev using one of its many >>>>> source adapters, and let it do the remote write >>>>> <https://vector.dev/docs/reference/configuration/sinks/prometheus_remote_write/> >>>>> for you. >>>>> >>>>> In option 2, what I was thinking was that you could expose the data as >>>>> a regular exporter, at the same time as you write it to MongoDB. >>>>> Prometheus can then scrape this as normal, i.e. it will scrape the *most >>>>> recent" updated value only. >>>>> >>>>> -- >>>>> >>>> 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/a84587cb-9db9-496c-a7f8-e1442c16a8e2n%40googlegroups.com >>>>> <https://groups.google.com/d/msgid/prometheus-users/a84587cb-9db9-496c-a7f8-e1442c16a8e2n%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/1e9affbe-96b4-4cdd-86e7-2b93543ab117n%40googlegroups.com >>> <https://groups.google.com/d/msgid/prometheus-users/1e9affbe-96b4-4cdd-86e7-2b93543ab117n%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/c1a6dff7-c0fc-4281-8293-fa666bccfc67n%40googlegroups.com > <https://groups.google.com/d/msgid/prometheus-users/c1a6dff7-c0fc-4281-8293-fa666bccfc67n%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/CAOBTUyVVNBCMm34oEdP0cEdNrHLYtoemPUb__BoHM2Gn0JLLUw%40mail.gmail.com.
