Monitoring & Metrics

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're using Prometheus Operator (for instance as part of kube-prometheus-stack, and deploying kuik with Helm, you can create PodMonitor and ServiceMonitor resources by setting the following values:

controllers.podMonitor.create=true
proxy.podMonitor.create=true
registry.serviceMonitor.create=true

And if you enabled minio to provide an S3-compatible storage for the registry, you can scrape its metrics as well with a ServiceMonitor by setting the following value:

minio.metrics.serviceMonitor.enabled=true

Exposed Metrics

Controller

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_containers_with_cached_image	Number of containers that have been rewritten to use a cached image
kube_image_keeper_controller_image_caching_request	Number of request to cache an image
kube_image_keeper_controller_image_put_in_cache_total	Count of all cached images since controller start
kube_image_keeper_controller_image_removed_from_cache_total	Count of all images removed from the cache since controller start
kube_image_keeper_controller_is_leader	Return 1 if the pod is leader
kube_image_keeper_controller_repositories	Number of repositories
kube_image_keeper_controller_up	Return 1 if the controller is running

By default, two replicas of the controller are running, and one of them becomes the leader. The value of cached_images should be the same across all replicas. However, the values for put_in_cache and removed_from_cache will increase only for the leader controller. They get reset to zero when the controller restarts, so they should mostly be used as "sign of life", or e.g. to detect when no images get removed from the cache even over multiple weeks or months.

Proxy

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

Registry

These metrics are exposed by the Docker Registry itself, more details in the official documentation

Grafana Dashboard

We provide a Grafana dashboard available in this repository or on GrafanaLabs.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!