Hi there,
Recently I’m trying to add a new Grafana Dashboard to the project https://github.com/slashdevops/kube-visibility, the dashboard is very simple, just show the version of all components installed into kubernetes.
my query is:
count by (namespace,job,version) ({__name__=~".*_build_info"})
My problem is that you are reporting two different Prometheus metrics grafana_build_info and grafana_plugin_build_info with the same labels values, what I mean the label service.
grafana_build_info{branch="HEAD",container="grafana",edition="oss",endpoint="3000",goversion="go1.15.1",instance="172.17.0.9:3000",job="grafana",namespace="kube-visibility",pod="grafana-5568cd6d99-n7gs7",revision="6668161a88",service="grafana",version="7.3.1"} 1
grafana_plugin_build_info{container="grafana",endpoint="3000",instance="172.17.0.9:3000",job="grafana",namespace="kube-visibility",plugin_id="input",plugin_type="datasource",pod="grafana-5568cd6d99-n7gs7",service="grafana",version="1.0.0"} 1
I know the metric name is different, but it is possible to report for the metric grafana_plugin_build_info a value service="grafana_plugin"? or wherever you want and different to the same label for metric grafana_build_info
or maybe a new label to differentiate these two components and future components like a new label component=grafana|plugins|new component. I know this is not recommended by Prometheus, but I only trying to help.
Thanks for this great product!
Christian