We have support for Kubernetes metrics from kubelet & most of kube-state-metrics, we should increase module coverage to also support Kubernetes internal services:

• apiserver Add kubernetes apiserver metricset #7059
• kube-proxy Add support to monitor kubernetes kube-proxy #10431
• kube-scheduler
• CoreDNS (1.11) Add support to monitor CoreDNS #10432
• kube-controller-manager
• etcd Support metrics endpoint in etcd metricbeat module #9093

Filebeat:

• Support /var/log/containers path for non-docker scenarios Input docker behaviour with Openshift 3.10 #8988

Metricbeat:

• Implement service account token based HTTP authorization Implement service account token based HTTP authorization #7518

Logs. Review where it makes sense to implement a module, for sure we want:

• Auditing

Metadata processor

• Include owner object reference info (replication controller, deployment, etc) Add owner object info to Kubernetes metadata #7231

• Seamless integrate it with kubernetes metricsets: Include labels and annotations in all Kubernetes metricsets #7148

As for kube-state-metrics:

• Resource Quotas [Metricbeat] support kubernetes ressourcequotas #6659
• ConfigMap
• Secret
• Daemonset
• Replication Controller
• Volume & Volume Claims, also check Add pvc name and namespace to kubernetes volume metricset #6977
• Job
• CronJob [Metricbeat] Cronjob support #12183
• Service
• HPA

Review:

• metricbeat state_container kube_pod_container_status_restarts -> kube_pod_container_status_restarts_total #6638

• Add kube_pod_container_status_<terminated/waiting>_reason to metricbeat/kubernetes/state_container #6639

Also we should consider joining these with kubelet sourced metrics, for total correlation: #6637