added support for optional clusterName via config file

[xmcqueen]

When there are multiple clusters in an area its essential to have the event's clustername associated with each event. The event object exposed by Kubernetes provides a clustername, but it does not set it. This PR sets the clustername.

#167

[xmcqueen]

The build is failing because it needs a mod tidy or something. The changes do not introduce any new dependencies. However, if I do that the PR becomes very noisy with unrelated vendor changes. Should I do a mod tidy and update the PR?

Maybe a 1 PR with nothing but mod tidy, and then a second one for this changeset would be easy to manage. A mod tidy is probably needed somewhere at this point.

[anannaya]

@xmcqueen Is it working for you. I built image from this PR , still i see empty string clustername.
below is my configuration

  config.yaml: |
    logLevel: debug
    logFormat: json
    throttlePeriod: 10
    ClusterName: staging-eks
    kubeQPS: 60
    kubeBurst: 60
    route:
      routes:
      - match:
        - receiver: "elasticsearch-dump"
    receivers:
    - name: "elasticsearch-dump"
      elasticsearch:
        hosts:
        - http://aws-es-log.staging.xxxxxx.com:80
        index: logstash-kube-events
        indexFormat: "logstash-kube-events-{2006-01-02}"
        deDot: true
        useEventID: true
        layout:
          component: '{{ .Source.Component }}'
          count: '{{ .Count }}'
          '@timestamp': '{{ .GetTimestampISO8601 }}'
          host: '{{ .Source.Host }}'
          clustername: '{{ .ClusterName }}'
          kind: '{{ .InvolvedObject.Kind }}'
          labels: '{{ toJson .InvolvedObject.Labels}}'
          message: '{{ .Message }}'
          name: '{{ .InvolvedObject.Name }}'
          namespace: '{{ .Namespace }}'
          reason: '{{ .Reason }}'
          type: '{{ .Type }}'

[xmcqueen]

Yes its in very heavy usage over here. It works great. I wonder if the problem is your use of ClusterName in the yaml instead of clusterName.

Here's the annotation for the struct where the yaml is decoded: pkg/exporter/config.go +16

	// TODO: I am not sure what to do here.
	LogLevel       string                    `yaml:"logLevel"`
	LogFormat      string                    `yaml:"logFormat"`
	ThrottlePeriod int64					 `yaml:"throttlePeriod"`
	ThrottlePeriod int64                     `yaml:"throttlePeriod"`
	ClusterName    string                    `yaml:"clusterName,omitempty"`
	Namespace      string                    `yaml:"namespace"`
	LeaderElection kube.LeaderElectionConfig `yaml:"leaderElection"`
	Route          Route    

[anannaya]

Yes its in very heavy usage over here. It works great. I wonder if the problem is your use of ClusterName in the yaml instead of clusterName.

Here's the annotation for the struct where the yaml is decoded: pkg/exporter/config.go +16

	// TODO: I am not sure what to do here.
	LogLevel       string                    `yaml:"logLevel"`
	LogFormat      string                    `yaml:"logFormat"`
	ThrottlePeriod int64					 `yaml:"throttlePeriod"`
	ThrottlePeriod int64                     `yaml:"throttlePeriod"`
	ClusterName    string                    `yaml:"clusterName,omitempty"`
	Namespace      string                    `yaml:"namespace"`
	LeaderElection kube.LeaderElectionConfig `yaml:"leaderElection"`
	Route          Route    

Thanks it works, Very sharp Eye.

[fpants]

@xmcqueen can I open a PR with this code for the maintained fork here? https://github.com/resmoio/kubernetes-event-exporter

[xmcqueen]

wow I had no idea. Yes and maybe I'll jump over there too. I've got 2 PRs queued up here now. Thanks

I need metrics in there. I've got a PR in here that adds prometheus metrics.

Wow and the vendor directory is gone too! I was going to open a ticket about that!

Yes, I'm jumping over to this new repo. You go ahead with the clusterName PR.

I'll get the prometheus metrics PR going.