You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Configmap metrics only export info, created, and resource_version (docs). It would be useful to see labels and annotations as well.
Why is this needed:
In my usecase, it's possible to use the labels/annotations applied to pods to track the rollout of versions of an application across different clusters (each with its own prometheus instance, aggregated in a single grafana dashboard -- the labels are applied by our deployment automation). Some of the "applications" have static (or rarely changing) code, and changes in behavior are instead driven by config map updates. It would be good to have a similar ability to track those rollouts via metrics, which requires exposing the labels and annotations to prometheus.
(In the above writeup I've referred to "labels and annotations". But perhaps it makes sense to prioritize one or the other -- that would probably work as well. Our tooling at the moment uses annotations but could probably be adjusted to use labels instead.)
Describe the solution you'd like
Exposure of kube_configmap_annotations and kube_configmap_labels -- exactly parallel to kube_namespace_annotationskube_pod_labels etc.
Additional context
The text was updated successfully, but these errors were encountered:
Hmm as far as I am aware in annotation there are some special fields (like lastAppliedConfiguration, generation etc) which should not be exposed (because of cardinality and sensitive information exposure)
What would you like to be added:
Configmap metrics only export info, created, and resource_version (docs). It would be useful to see labels and annotations as well.
Why is this needed:
In my usecase, it's possible to use the labels/annotations applied to pods to track the rollout of versions of an application across different clusters (each with its own prometheus instance, aggregated in a single grafana dashboard -- the labels are applied by our deployment automation). Some of the "applications" have static (or rarely changing) code, and changes in behavior are instead driven by config map updates. It would be good to have a similar ability to track those rollouts via metrics, which requires exposing the labels and annotations to prometheus.
(In the above writeup I've referred to "labels and annotations". But perhaps it makes sense to prioritize one or the other -- that would probably work as well. Our tooling at the moment uses annotations but could probably be adjusted to use labels instead.)
Describe the solution you'd like
Exposure of
kube_configmap_annotations
andkube_configmap_labels
-- exactly parallel tokube_namespace_annotations
kube_pod_labels
etc.Additional context
The text was updated successfully, but these errors were encountered: