Logging on CoreOS (GCE with Containers Instance Group) produces errors/warnings

[JanMikes]

Hello,

We are using managed instance groups on Google Compute Engine with containers, os is powered by coreOS so stackdriver monitoring agent can not be installed directly to the host.
This way i have discovered your docker image and have tried it.

For me it seems that it sends metrics correctly to the stackdriver, but it produces a lot of warning/error output, like this:

[2019-09-24 10:07:58] write_gcm: Server response (CollectdTimeseriesRequest) contains errors:
{
  "payloadErrors": [
    {
      "index": 4,
      "error": {
        "code": 3,
        "message": "Expected 4 labels. Found 0. Mismatched labels for payload [values {\n  data_source_name: \"read\"\n  data_source_type: DERIVE\n  value {\n    int64_value: 0\n  }\n}\nvalues {\n  data_source_name: \"write\"\n  data_source_type: DERIVE\n  value {\n    int64_value: 1458176\n  }\n}\nstart_time {\n  seconds: 1569319197\n  nanos: 586014582\n}\nend_time {\n  seconds: 1569319617\n  nanos: 572968679\n}\nplugin: \"processes\"\nplugin_instance: \"all\"\ntype: \"disk_octets\"\n] on resource [type: \"gce_instance\"\nlabels {\n  key: \"instance_id\"\n  value: \"2178816110945768016\"\n}\nlabels {\n  key: \"zone\"\n  value: \"europe-west1-c\"\n}\n] for project 796672424361"
      }
    },
    {
      "index": 145,
      "error": {
        "code": 3,
        "message": "Expected 4 labels. Found 0. Mismatched labels for payload [values {\n  data_source_n
[2019-09-24 10:07:59] write_gcm: Server response (CollectdTimeseriesRequest) contains errors:
{
  "payloadErrors": [
    {
      "index": 4,
      "error": {
        "code": 3,
        "message": "Expected 4 labels. Found 0. Mismatched labels for payload [values {\n  data_source_name: \"read\"\n  data_source_type: DERIVE\n  value {\n    int64_value: 0\n  }\n}\nvalues {\n  data_source_name: \"write\"\n  data_source_type: DERIVE\n  value {\n    int64_value: 1544192\n  }\n}\nstart_time {\n  seconds: 1569319197\n  nanos: 586014582\n}\nend_time {\n  seconds: 1569319677\n  nanos: 571233591\n}\nplugin: \"processes\"\nplugin_instance: \"all\"\ntype: \"disk_octets\"\n] on resource [type: \"gce_instance\"\nlabels {\n  key: \"instance_id\"\n  value: \"2178816110945768016\"\n}\nlabels {\n  key: \"zone\"\n  value: \"europe-west1-c\"\n}\n] for project 796672424361"
      }
    }
  ]
}

I have started the container with this command

sudo docker run -e MONITOR_HOST=true -v /proc:/mnt/proc:ro --privileged wikiwi/stackdriver-agent

Do you have any idea how to fix this?