Files

Paul Payne 9687fad812 feat: Move cluster services to wild-directory as unified packages

Convert all 15 cluster services from embedded API format to
wild-directory packages using the unified manifest format:
- metallb, traefik, cert-manager, longhorn, snapshot-controller
- nfs, smtp, coredns, node-feature-discovery, nvidia-device-plugin
- externaldns, docker-registry, headlamp, crowdsec, utils

Changes:
- wild-manifest.yaml → manifest.yaml with is, defaultConfig, requires
- Eliminated configReferences and serviceConfig fields
- Flattened kustomize.template/ to package root
- Template vars use flat defaultConfig keys
- install.sh paths updated for apps/ layout
- Updated 9 app manifests: cloud.smtp.* → apps.smtp.* with requires
- Removed dead install: true field from 6 app manifests

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2026-05-17 02:26:46 +00:00

1.3 KiB

Raw Blame History

CoreDNS

CoreDNS has the kubernetes plugin, so it returns all k8s service endpoints in well-known format.

All services and pods are registered in CoreDNS.

..svc.cluster.local
.
(if in the same namespace)
..pod.cluster.local
...svc.cluster.local

Any query for a resource in the internal.$DOMAIN domain will be given the IP of the Traefik proxy. We expose the CoreDNS server in the LAN via MetalLB just for this capability.

Default CoreDNS Configuration

This is the default CoreDNS configuration, for reference:

.:53 {
    errors
    health { lameduck 5s }
    ready
    log . { class error }
    prometheus :9153
    kubernetes cluster.local in-addr.arpa ip6.arpa {
        pods insecure
        fallthrough in-addr.arpa ip6.arpa
        ttl 30
    }
    forward . /etc/resolv.conf { max_concurrent 1000 }
    cache 30 {
        disable success cluster.local
        disable denial cluster.local
    }
    loop
    reload
    loadbalance
}

1.3 KiB Raw Blame History

CoreDNS

Default CoreDNS Configuration

1.3 KiB

Raw Blame History