Glossary¶

2026-04-27

Defined here once, referenced consistently throughout the rest of this site.

Alertmanager¶

Component that receives alerts from one or more Prometheis, deduplicates and groups them, and sends notifications via receivers (Slack, PagerDuty, email, ...). Configured via the Alertmanager CR + alertmanager.config values + AlertmanagerConfig CRs.

AlertmanagerConfig¶

Namespaced Custom Resource that lets a team define routing and receivers without editing the chart's alertmanager.config. Useful for delegating alert ownership to product teams.

Chart umbrella / subchart¶

This chart is an umbrella chart. It declares dependencies on other charts (its subcharts: crds, kube-state-metrics, prometheus-node-exporter, grafana, prometheus-windows-exporter) in Chart.yaml and bundles them into a single deploy.

CRD (CustomResourceDefinition)¶

The schema for a Custom Resource. The Operator owns 10 CRDs in the monitoring.coreos.com API group: Prometheus, Alertmanager, AlertmanagerConfig, PrometheusRule, ServiceMonitor, PodMonitor, Probe, ScrapeConfig, PrometheusAgent, ThanosRuler.

CR (Custom Resource)¶

An instance of a CRD. The chart creates Prometheus, Alertmanager, PrometheusRule, and ServiceMonitor CRs. The Operator turns those CRs into StatefulSets, Secrets, and config.

External labels¶

Labels added to every metric leaving Prometheus (prometheus.prometheusSpec.externalLabels). Used to identify the source cluster / environment in federated and remote-write setups.

Grafana¶

Dashboards UI. Deployed via the grafana subchart. Pre-configured with Prometheus as the default datasource and a sidecar that auto-loads dashboards from any ConfigMap labelled grafana_dashboard: "1".

kube-prometheus-stack¶

The upstream chart name (under prometheus-community/helm-charts). The OVES copy lives at deployment-charts/prometheus-stack/.

kube-state-metrics¶

Exporter that exposes metrics about Kubernetes objects (Deployments, Pods, Nodes, ...). Distinct from node-exporter, which exposes node-level OS metrics.

node-exporter¶

DaemonSet exporter that exposes Linux host metrics from each node (CPU, memory, disk, network, file system).

PodMonitor¶

Custom Resource describing a scrape job that targets pods directly (not through a Service). Useful when there is no Service, or when you want per-pod-IP granularity.

Prometheus¶

The TSDB and scrape engine.
The Custom Resource (apiVersion: monitoring.coreos.com/v1) that the Operator reconciles into a Prometheus StatefulSet.

Prometheus Operator¶

The Kubernetes controller that watches the chart's CRDs and materializes them as workloads. Source of truth for everything in the chart's own templates.

PrometheusRule¶

Custom Resource containing alerting and recording rules. Picked up by the Operator if its labels match prometheus.prometheusSpec.ruleSelector.

Probe¶

Custom Resource for synthetic black-box checks (HTTP / TCP / ICMP) executed by prometheus-blackbox-exporter. Not bundled in this chart; install blackbox-exporter separately.

Receiver¶

Alertmanager term for a notification destination (Slack channel, PagerDuty service, webhook URL, email address, ...). Defined under alertmanager.config.receivers.

Recording rule¶

A precomputed PromQL expression saved as a new metric. Reduces query cost for expensive expressions used by dashboards.

ScrapeConfig¶

Newer Custom Resource (alpha in older Operator releases) for arbitrary scrape jobs that don't fit ServiceMonitor / PodMonitor, e.g. EC2 service-discovery, file-based SD, custom static configs.

ServiceMonitor¶

Custom Resource describing a scrape job that targets a Kubernetes Service. The most common way to add scrape targets.

Sharding¶

Splitting Prometheus by series across multiple Prometheus instances using prometheus.prometheusSpec.shards. Different from replicas, which run the same scrape config on multiple instances for HA. Sharding requires a front-end (Thanos / Cortex / Mimir) to query across shards.

Thanos¶

Long-term, deduplicated, globally-queryable Prometheus. Out of scope for this chart; the chart can deploy a ThanosRuler CR for centralized rule evaluation, but the rest of Thanos must be installed separately.

Validating admission webhook¶

The Operator runs a validating webhook that rejects malformed PrometheusRule resources at kubectl apply time. Toggled by prometheusOperator.admissionWebhooks.enabled.