Kube-proxy-metrics
It is highly recommended gathering metrics from kube-proxy as it is a critical piece of any Kubernetes Cluster. Sometimes (especially in managed clusters) it is not possible to configure kube-proxy to expose metrics, this is why this package exists. Another reason to run this exporter instead of just exposing metrics from kube-proxy is the ability to run it independently of the environment, on-premise installed by kubeadm or a managed Kubernetes Cluster.
Requirements
- Kubernetes >=
1.20.0
- Kustomize =
v3.3.X
- prometheus-operator
Image repository and tag
- kube-rbac-proxy image:
registry.sighup.io/fury/brancz/kube-rbac-proxy:v0.11.0
- kube-rbac-proxy repository: kube-rbac-proxy on Github
Configuration
Fury distribution kube-proxy-metrics is deployed with the following configuration:
- Resource limits are
20m
for CPU and40Mi
for memory - Listens on port
18443
- Metrics are scraped by Prometheus with
15s
intervals - Requires
hostNetwork: true
andhostPID: true
- Runs as non-root user.
Deployment
You can deploy kube-proxy-metrics by running the following command in the root of the project:
kustomize build | kubectl apply -f -