Default Monitoring Components

The monitoring stack deploys the following components in the environment to provide a complete solution for monitoring the infrastructure, receiving alerts, and consulting performance graphs.

  • Cluster Monitoring Operator
    The cluster monitoring operator (CMO) is the central component of the monitoring stack. It controls the deployed monitoring components and ensures that they are always in sync with the latest version of the CMO.

  • Prometheus Operator
    The Prometheus operator deploys and configures both Prometheus and Alertmanager. The operator also manages the generation and configuration of configuration targets (service monitors and pod monitors).

  • Prometheus
    Prometheus is the monitoring server.

  • Prometheus Adapter
    The Prometheus adapter exposes cluster resources that are used for Horizontal Pod Autoscaling (HPA).

  • Alertmanager
    Alertmanager handles alerts sent by the Prometheus server.

  • kube-state-metrics
    kube-state-metrics is a converter agent that exports Kubernetes objects to metrics that Prometheus can parse.

  • openshift-state-metrics
    openshift-state-metrics is based on the kube-state-metrics and adds monitoring for OKD-specific resources (such as image registry metrics).

  • node-exporter
    node-exporter exports low-level metrics for worker nodes.

  • Thanos Querier
    Thanos Querier is a single, multi-tenant interface that enables aggregating and deduplicating cluster and user workload metrics.

  • Grafana
    Grafana is a platform for analyzing and visualizing metrics. The Grafana dashboards provided with the monitoring stack are read-only

  • Telemeter Client
    The Telemeter Client sends a subsection of the data from platform Prometheus instances to Red Hat to facilitate Remote Health Monitoring for clusters.

All of the components in the monitoring stack are monitored by the stack and are automatically updated when OKD is updated.