To assist with operations and provide a visualization of kube-image-keeper activities, Prometheus metrics are exposed from the three components (proxy, controller and registry).
If you are deploying kube-image-keeper with Helm, both PodMonitor and ServiceMonitor can be configured through the following values:
controllers.podMonitor.create=true
proxy.podMonitor.create=true
registry.serviceMonitor.create=true
If you use minio as a S3 compatible storage for the registry, you should be able to get metrics by enabling a serviceMonitor for minio:
minio.metrics.serviceMonitor.enabled=true
Metric | Description |
---|---|
kube_image_keeper_controller_build_info | Provide informations about controller version |
kube_image_keeper_controller_cached_images | Count of all cached images expired or not |
kube_image_keeper_controller_image_put_in_cache_total | Count of all cached images since the bootstrap |
kube_image_keeper_controller_image_removed_from_cache_total | Count of all images removed from the cache |
kube_image_keeper_controller_is_leader | Return 1 if the pod is leader |
kube_image_keeper_controller_up | Return 1 if the controller is running |
Metric | Description |
---|---|
kube_image_keeper_proxy_build_info | Provide informations about proxy version |
kube_image_keeper_proxy_http_requests_total | Provide information about cache hit and http requests |
These metrics are exposed by the Docker Registry itself, more details in the official documentation
We provide a Grafana dashboard available here or directly on GrafanaLabs.