Prometheus exporter for cluster-autoscaler-status

k8s provides a cluster-autoscaler-status configmap. sadly there was no way to easily export this data to Prometheus.
now there is.

we export three metrics per pool:

• number of ready nodes
• cloudProviderTarget number
• maxSize number

Example output

# HELP cluster_autoscaler_pool ready
# TYPE cluster_autoscaler_pool gauge
cluster_autoscaler_pool{name="csi"} 0.0
cluster_autoscaler_pool{name="engines"} 1.0
# HELP cluster_autoscaler_pool_target cloudProviderTarget
# TYPE cluster_autoscaler_pool_target gauge
cluster_autoscaler_pool_target{name="csi"} 0.0
cluster_autoscaler_pool_target{name="engines"} 1.0
# HELP cluster_autoscaler_pool_max maxSize
# TYPE cluster_autoscaler_pool_max gauge
cluster_autoscaler_pool_max{name="csi"} 10.0
cluster_autoscaler_pool_max{name="engines"} 50.0

Usage

it can be run in-cluster or out of cluster. metrics exporter runs on port 9100

there's a ready to use Docker image

Limitations

currently the querying interval is hardcoded at 10s