Helm values

The Helm chart exposes image repositories and tags for the operator and runtime components, plus standard Kubernetes settings such as ServiceAccount, RBAC, resources, node selectors, tolerations, and affinity.

The chart pins tested runtime images. Override image values only when testing a specific component build.

Operator failover HA is opt-in:

operator:
  replicas: 3
  highAvailability:
    enabled: true
    peeringName: ""

When HA is enabled and peeringName is empty, the chart uses the release fullname. Kopf peering keeps the operator active/passive: one pod reconciles while standby pods wait for failover, provided generated priorities are unique. The chart derives priorities from pod IPs, so they are probabilistically unique and intended for small replica counts, normally 2 or 3.

This is controller failover HA only. It does not make Kamailio, RTPengine, Asterisk, PostgreSQL, or SIP/RTP traffic handling highly available. The Kopf peering CRDs are cluster-scoped resources shared by all KubeVoIP releases, even though each release creates a namespaced KopfPeering.