Export more metric labels

[nambrot]

Howdy, I was wondering if it is possible to configure the sidecar to export some specific labels of our metrics? Or are we constrained to the generic task resource labels in Stackdriver?

[igorpeshansky]

The sidecar supports a number of resources, all of which are listed in resource_map.go. Most of the other labels are attached to the timeseries, except for the target labels and __name__ (see series_cache.go).
Does this answer your question?

[nambrot]

Thank you for your prompt response! I'm not entirely sure I understand the Resource Map. When we have run it on our cluster, it's been translated to a Generic Task. If I understand correctly, if I have some application specific metrics, that will be my best shot and using job and task_id as the only free labels I could use for my purposes?

[nambrot]

The reason I ask is that with prometheus-to-sd I was able to get all the labels from my prometheus metrics into stackdriver as well

[nambrot]

@igorpeshansky Just wondering if you had any updates/tips here?

[WaterKnight1998]

@nambrot Same happening here. If I mount prometheus-to-sd I get metrics (kube-state-metrics & Node exporter) from all namespaces in GCP Monitoring. However, If I use this sidecar I don´t get them from all namespaces.

How did you solve this error?

[clguo-tw]

same problem, can't get metrics from all namespaces
only get metrics from namespace which has prometheus