kube-prometheus is a Kubernetes monitoring stack that combines Prometheus, Alertmanager, Grafana, and related exporters with pre-configured dashboards and alerting rules.
The project solves the problem of setting up comprehensive cluster monitoring on Kubernetes by bundling together the Prometheus Operator, a highly available Prometheus and Alertmanager deployment, node-exporter, blackbox-exporter, kube-state-metrics, the Prometheus Adapter for Kubernetes metrics APIs, and Grafana. The entire stack is defined in Jsonnet, allowing it to function both as a ready-to-deploy package and as a customizable library. It comes pre-configured to collect metrics from all Kubernetes components and includes a default set of dashboards and alerting rules, many sourced from the kubernetes-mixin project.
Teams should adopt this tool if they need end-to-end cluster monitoring without building a monitoring stack from scratch. It suits organizations running Kubernetes who want both infrastructure and application monitoring with sensible defaults. The project requires only a Kubernetes cluster as a prerequisite, though it does assume the kubelet is configured with token authentication and authorization rather than client certificates, which enables more fine-grained access control. The stack provides resource metrics through the Prometheus Adapter, which requires Kubernetes to have the Extension API Server feature enabled.
The project carries an explicit experimental warning in its README, indicating that significant changes may occur at any time. Testing in CI covers only the last two Kubernetes releases and the main branch on a regular basis, though other versions may work. The codebase is written in Jsonnet, which means users who want to customize the deployment need familiarity with that language or willingness to learn it.
