Hi,
Thanks for your answer Christian.
Your idea was good, I tried to remove all unused labels and finaly succeed
on getting my metric joined correctly.
I used without because I see it in an answer of Brian Brazil and it did the
job.
Final aggregation
sum
(container_memory_rss{container_label_com_docker_swarm_task_name=~".+"})
without (instance, job) * on(container_label_com_docker_swarm_node_id)
group_left(node_name) node_meta{job="node-exporter"}
Thank you, it resolved
Have a good day
Le mardi 10 mars 2020 23:33:33 UTC+1, BDT a écrit :
>
> Hi everyone,
>
> Today I have a problem about my rules expression because I try to join
> metrics together to get the name of the swam node in it.
> In order to do this, I have left_joined my metrics by node_id and get
> node_name. It's works fine in the prometheus console but when I deploy my
> rules, I get this message:
>
> level=warn ts=2020-03-10T20:27:12.111Z caller=manager.go:525
> component="rule manager" group="Container alert" msg="Evaluating rule
> failed"
>
> rule="alert: task_high_memory_usage_1g
>
> expr: sum by(container_label_com_docker_swarm_task_name,
> container_label_com_docker_swarm_node_id)
> (container_memory_rss{container_label_com_docker_swarm_task_name=~".+"}) *
> on(container_label_com_docker_swarm_node_id) group_left(node_name)
> node_meta > 1e+06
>
> err="found duplicate series for the match group
> {container_label_com_docker_swarm_node_id=\"jilcpclonjg7jj1chh29b19pl\"} on
> the right hand-side of the operation:
>
> [{__name__=\"node_meta\",
> container_label_com_docker_swarm_node_id=\"jilcpclonjg7jj1chh29b19pl\",
> instance=\"10.0.9.10:9100\", job=\"node-exporter\",
> node_id=\"jilcpclonjg7jj1chh29b19pl\", node_name=\"***\"},
>
> {__name__=\"node_meta\",
> container_label_com_docker_swarm_node_id=\"jilcpclonjg7jj1chh29b19pl\",
> instance=\"10.0.8.20:9100\", job=\"node-exporter\",
> node_id=\"jilcpclonjg7jj1chh29b19pl\", node_name=\"***\"}];
>
> many-to-many matching not allowed: matching labels must be unique on one
> side"
>
> I have a label instance with two different values. my first metric is get
> by cadvisor and the other is generated by node_exporter so two different
> instances
>
> node_meta{container_label_com_docker_swarm_node_id="jilcpclonjg7jj1chh29b19pl",
>
> instance="10.0.9.10:9100", job="node-exporter",
> node_id="jilcpclonjg7jj1chh29b19pl", node_name="***"}
>
> This is an example of container_memory_rss:
>
>
> container_memory_rss{container_label_caddy_version="1.0.3",container_label_com_docker_stack_namespace="supervision",container_label_com_docker_swarm_node_id="jilcpclonjg7jj1chh29b19pl",container_label_com_docker_swarm_service_id="5t7qdfsjdirecjvw0nlpzlw0w",container_label_com_docker_swarm_service_name="supervision_dockerd-exporter",container_label_com_docker_swarm_task_id="n8ji42apbj0t2bctr6ilgcvt2",container_label_com_docker_swarm_task_name="supervision_dockerd-exporter.jilcpclonjg7jj1chh29b19pl.n8ji42apbj0t2bctr6ilgcvt2",",id="/docker/312f83a134e144e3342d6229ec1a239e0539b77ec79f3a7e493f04fc9e24edf1",image="
> registry.test.toto.fr/dockerd-exporter:1.0.3@sha256:31f18339414875c404647acac452883e2a7d6fdf89b8d8c8ae5e61140c392b80
> ",instance="10.0.9.20:8080
> ",job="cadvisor",name="supervision_dockerd-exporter.jilcpclonjg7jj1chh29b19pl.n8ji42apbj0t2bctr6ilgcvt2"}
>
> The alert is sent anyway but I don't like have errors in app ^^
> Maybe there is a way to delete instance label before doing the left_join,
> I don't need this label in my expression.
>
> Thanks for your time !
>
> Best regards
>
--
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/aee5ba28-55a2-4b36-8a72-4c4ce90f70bd%40googlegroups.com.
