Well, "many" means "one or more", and therefore "one" is a valid case of "many" :-)
But yes, you'll have to treat it as a many-to-one in order to pick up the extra labels. If you know for sure that metrics A and B are always matched 1:1, then it doesn't matter whether you use group_left or group_right. But you still need to provide the list of labels in brackets in order to pick up the labels from the other side. On Tuesday, 7 March 2023 at 12:43:22 UTC Nick Carlton wrote: > Ah ok, > > I would have thought this was one-to-one because for each metric that > exists in ltmPoolMbrStatusAvailState there is one exact match within > ltmPoolMbrStatusEnabledState. Not multiple. > > Unless I’m reading the definitions wrong. > > Thanks > > On Tue, 7 Mar 2023 at 12:06, Brian Candler <[email protected]> wrote: > >> To pick up labels from the other side, you need to list them as part of >> your group_right. e.g. (untested) >> >> ltmPoolMbrStatusAvailState * >> on(instance,ltmPoolMbrStatusNodeName,ltmPoolMbrStatusPoolName,ltmPoolMbrStatusPort) >> >> group_right(*foo,bar,baz*) ltmPoolMbrStatusEnabledState >> >> will pick up labels foo,bar,baz from the left side. >> >> See: >> https://prometheus.io/docs/prometheus/latest/querying/operators/#many-to-one-and-one-to-many-vector-matches >> >> Here's a tested example: >> >> node_filesystem_avail_bytes * on (instance) >> group_left(machine,release,version) node_uname_info >> >> It's a many-to-one, where the left side is "many" and the right side is >> "one", and the given labels from node_uname_info are added to the labels >> from node_filesystem_avail_bytes. >> >> On Tuesday, 7 March 2023 at 08:58:21 UTC Nick Carlton wrote: >> >>> Thanks for your response. >>> >>> If I do `ltmPoolMbrStatusAvailState * >>> on(instance,ltmPoolMbrStatusNodeName,ltmPoolMbrStatusPoolName,ltmPoolMbrStatusPort) >>> >>> group_right ltmPoolMbrStatusEnabledState` I get: >>> >>> `{instance="ltm01", job="f5_ltm_test", ltmPoolMbrStatusEnabledState="1", >>> ltmPoolMbrStatusNodeName="/Common/VPN1", >>> ltmPoolMbrStatusPoolName="/Common/Pool1", ltmPoolMbrStatusPort="4500", >>> prometheus="monitoring/k8s", prometheus_replica="prometheus-k8s-0"}` >>> >>> So this has `ltmPoolMbrStatusEnabledState` but not >>> `ltmPoolMbrStatusAvailState`. If I swap the metrics round: >>> >>> `ltmPoolMbrStatusEnabledState * >>> on(instance,ltmPoolMbrStatusNodeName,ltmPoolMbrStatusPoolName,ltmPoolMbrStatusPort) >>> >>> group_right ltmPoolMbrStatusAvailState`. I get: >>> >>> `{instance="ltm01", job="f5_ltm_test", ltmPoolMbrStatusAvailState="1", >>> ltmPoolMbrStatusNodeName="/Common/VPN1", >>> ltmPoolMbrStatusPoolName="/Common/Pool1", ltmPoolMbrStatusPort="4500", >>> prometheus="monitoring/k8s", prometheus_replica="prometheus-k8s-0"}` >>> >>> This has `ltmPoolMbrStatusAvailState` but not >>> `ltmPoolMbrStatusEnabledState` >>> >>> So it seems I can get one or the other, not both. Unless im forming my >>> PromQL wrong? >>> >>> On Tuesday, 7 March 2023 at 08:31:39 UTC Brian Candler wrote: >>> >>>> It has to be done in a query - the relabelling phase of a scrape job >>>> cannot see other metrics. >>>> >>>> What you are looking for is one-to-many queries, which can pick up >>>> labels from the "one" side and apply them to the "many": >>>> >>>> https://www.robustperception.io/how-to-have-labels-for-machine-roles >>>> >>>> https://www.robustperception.io/exposing-the-software-version-to-prometheus >>>> >>>> https://prometheus.io/docs/prometheus/latest/querying/operators/#many-to-one-and-one-to-many-vector-matches >>>> https://www.robustperception.io/left-joins-in-promql >>>> >>>> > I have tried doing a join but it only seems to bring in one of the >>>> `ltmPoolMbrStatusEnabledState` and `ltmPoolMbrStatusAvailState` metrics >>>> depending on if its a left or right join, but I would like both to appear. >>>> >>>> Can you show what query you did, the raw metrics which it used, what >>>> result you got, and what you wanted to get instead? >>>> >>> -- >> You received this message because you are subscribed to a topic in the >> Google Groups "Prometheus Users" group. >> To unsubscribe from this topic, visit >> https://groups.google.com/d/topic/prometheus-users/qlHK7r1hQVs/unsubscribe >> . >> To unsubscribe from this group and all its topics, send an email to >> [email protected]. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/prometheus-users/4c22a279-a51c-47db-993e-9107e54e883en%40googlegroups.com >> >> <https://groups.google.com/d/msgid/prometheus-users/4c22a279-a51c-47db-993e-9107e54e883en%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/cb465a72-3dff-436c-b0da-c437193ceb8en%40googlegroups.com.
