This repository has been archived by the owner on Oct 3, 2023. It is now read-only.
Missing Zone label #14
Comments
rghetia
pushed a commit
to census-ecosystem/opencensus-go-exporter-stackdriver
that referenced
this issue
Dec 3, 2019
* Use standard label name for location. The location label is required. The resource type mappings defined in resource.go uniformly expect resourcekeys.CloudKeyZone, except the `generic_task`, but there is no good reason for that, as Stackdriver still requires a valid cloud location for `generic_task`. Therefore, use a cloud location everywhere. This makes the exporter work out-of-the-box in conjuction with `kubernetes-operator` and partially addresses census-ecosystem/kubernetes-operator#14 * Update stackdriver.go Added printfs that are useful for diagnosing resource discovery.
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Kube operator injects following resources labels with resource Type as
k8s.opencensus-clusterdefaultopencensus-pod-autoconfopencensusThis gets translated to 'k8s_container` resource in Stackdriver exporter as per this mapping
The problem here is that there is no Zone label injected by kube operator. As a result there is no
locationlabel after conversion in SD Exporter. 'location' being a mandatory label for 'k8s_container' resource type, the timeseries gets rejected due to missinglocationlabel.This could be designated as SD Exporter issue but kube operator becomes useless for SD Exporter without location. Anyone who wants to use these resource labels would have to add gke.Detect() in SD exporter option . Hence the issue is opened here to add Zone label support.
The text was updated successfully, but these errors were encountered: