The relabelling process doesn't happen on the *up *metric, so the metric doesn't have the correct labels I need for my query.
To illustrate this, imagine I have two services generating metrics - service A generates metrics with a label svc="service-a" - service B generates metrics with a label svc="service-b" For whatever reason I want to relabel *svc* to *service*. This works for all the metrics other than *up*. I want to run Prometheus queries against different metrics using a consistent set of labels (e.g. *service *rather than *svc*). This is especially important using visualisation tools like Grafana (imagine a *service *drop down box). On Wednesday, 13 January 2021 at 14:13:27 UTC Stuart Clark wrote: > On 2021-01-13 13:55, Tom Gregory wrote: > > I'm using Prometheus federation i.e. I'm ingesting metrics from one > > Prometheus to another. > > > > To do some relabelling I'm using _METRIC_RELABEL_CONFIGS_. According > > to the docs [1]: > > > > METRIC RELABELLING DOES NOT APPLY TO AUTOMATICALLY GENERATED > > TIMESERIES SUCH AS UP. > > > > That makes me sad because I want to measure uptime using these > > metrics. Is there any suggested approach to getting around this > > problem please? Any way to measure uptime without the automatically > > generated _up_ metric? > > > > What is the problem you are having with the up metric? > > up (and a few others) are created by Prometheus to give some details > about scrapes. So they should "just work". > > -- > Stuart Clark > -- 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/2b55e54f-d0cb-493d-ae69-0ef02bcf4edcn%40googlegroups.com.
