Getting kube-prometheus-stack running starts with three things: a Kubernetes cluster on v1.19 or later, Helm 3.x, and a dedicated namespace for monitoring workloads. The installation itself is a single Helm command, but getting it production-ready involves a handful of decisions worth making up front — which namespace to use, whether to pin a chart version, and what to verify once the pods come up.
This category covers the installation path end to end: adding the prometheus-community Helm repository, running the install command, and confirming that Prometheus, Grafana, Alertmanager, and the supporting exporters all reach a healthy state. It is the natural starting point before moving on to configuration or production hardening.
What actually changes when you install kube-prometheus-stack on k3s, RKE2, Amazon EKS, minikube, microk8s, and OpenShift — storage classes, ingress controllers, and security contexts.
Read ArticleThe most complete kube-prometheus-stack Helm chart documentation for 2026. Installation, values.yaml configuration, CRDs, upgrading, production best practices — everything in one place.
Read ArticleA Kubernetes cluster running v1.19 or later, Helm 3.x installed locally, and kubectl configured with cluster access. A default StorageClass is recommended if you plan to enable persistence from the start.
Most teams use a dedicated monitoring namespace created with kubectl create namespace monitoring, keeping monitoring workloads isolated from application namespaces for RBAC and resource-quota purposes.
