Skip to content

Initial release of the Tenzir Node Helm chart.

Optional NetworkPolicy and PodDisruptionBudget

Section titled “Optional NetworkPolicy and PodDisruptionBudget”

Jun 17, 2026 · @Zedoraps, @claude · #2

Render an optional NetworkPolicy scoping ingress to the node pods by setting networkPolicy.enabled: true. Render an optional PodDisruptionBudget spanning every node pod in the release by setting podDisruptionBudget.minAvailable or podDisruptionBudget.maxUnavailable. Neither resource is rendered with the chart defaults.

Jun 17, 2026 · @Zedoraps, @claude · #2

Container securityContext ships hardened by default: runAsNonRoot: true with runAsUser and runAsGroup pinned to 999 (matching the image’s tenzir user), seccompProfile: { type: RuntimeDefault }, allowPrivilegeEscalation: false, capabilities.drop: [ALL], and readOnlyRootFilesystem: true. An emptyDir mounted at /tmp keeps the writable surface to just that path and the /var/lib/tenzir PVC.

Jun 17, 2026 · @Zedoraps, @claude · #2

Open additional listener ports through nodes[].extraPorts, which attaches a port to one specific node’s pod and Service (optionally backed by its own dedicated Service when serviceType is set), or through sharedServices, which creates a single fleet-wide Service whose endpoints span every selected node’s pod so kube-proxy load-balances across them.

Per-node configuration with checksum-driven rollouts

Section titled “Per-node configuration with checksum-driven rollouts”

Jun 17, 2026 · @Zedoraps, @claude · #2

Compose each node’s tenzir.yaml from a global tenzir.config overlay merged with each entry’s nodes[].config. A per-node checksum/config annotation on the pod template ensures helm upgrade only rolls the pods whose merged configuration actually changed; untouched nodes keep running.

Jun 17, 2026 · @Zedoraps, @claude · #2

Install the chart from oci://ghcr.io/tenzir/charts/tenzir-node to deploy one or more tenzir-node instances on Kubernetes. Each entry in the chart’s nodes list renders as its own one-pod StatefulSet with a dedicated Service, ConfigMap, Secret, and persistent volume claim, scaled independently of the others.