per elastic/apm#21
This logic we should follow is described in https://github.com/elastic/apm/blob/master/docs/agent-development.md#containerkubernetes-metadata
/proc/self/cgroup is not yet parsed. https://github.com/elastic/apm-agent-dotnet/pull/741#pullrequestreview-402115843
_zube 馃槧_
馃嵖
Is this me against @graphaelli and zube? 馃ぃ
So... this issue is almost done - @vhatsura made big progress on it - we already read and send environment variables. But /proc/self/cgroup is not yet parsed - so let's keep it open for that missing part.
Hi,
I'm still trying to understand what this and all related issues are about, but does this issue have anything to do with the integration between APM and logs? Or is this a complete different topic?
The issue I'm having is, when a dotnet application runs on kubernetes I see the container.id in the metadata.
When I choose the action "Pod Logs" or "Pod Metrics" it sends the container id to be used with kubernetes.pod.uid, which doesn't work of course.
I tested this with java and nodejs applications in APM and for them it uses the correct kubernetes.pod.uid.
So, I was wondering if this is an elastic/kibana bug or does this have to do with the agent.
For the latter, I can create a new issue of this is off topic of course.
Thanks,
Peter
I can confirm that indeed container.id is sent as kubernetes.pod.uid as well (PR #935 is based on 1.6.1). See https://apm-ci.elastic.co/job/apm-agent-dotnet/job/apm-agent-dotnet-mbp/job/PR-935/1/artifact/target/diag-Elastic.Apm.DockerTests.csproj.log:
systemInfo.Container.Id: 2227daf62df6694645fee5df53c1f91271546a9560e8600a525690ae252b7f63
systemInfo.Kubernetes.Pod.Uid: 2227daf62df6694645fee5df53c1f91271546a9560e8600a525690ae252b7f63
systemInfo.Container.Id: 15aa6e53-b09a-40c7-8558-c6c31e36c88a
systemInfo.Kubernetes.Pod.Uid: 15aa6e53-b09a-40c7-8558-c6c31e36c88a
cgroup is now also parsed - this is complete.