## Summary - Add layer4 TCP proxy configuration to Caddyfile template for SSH services - Configure Forgejo SSH on port 2222 → localhost:2200 - Switch HTTPS from port 8443 (testing) to 443 (production) - Requires Caddy rebuilt with `github.com/mholt/caddy-l4` plugin ## What This Enables Git+SSH access via `forge.ops.eblu.me:2222` is now accessible from: - Tailnet clients (gilbert) - Docker containers on indri - Kubernetes pods in minikube This solves the DNS resolution issues where containers couldn't reach Tailscale MagicDNS names. ## Testing Done - [x] Caddy rebuilt with layer4 plugin - [x] Validated Caddyfile syntax - [x] Cleared `svc:forge` from tailscale serve - [x] Verified HTTPS works: `curl https://forge.ops.eblu.me` - [x] Verified SSH works: `ssh -p 2222 forgejo@forge.ops.eblu.me` - [x] Verified git clone works via new endpoint - [x] Verified minikube pods can reach both HTTPS and SSH endpoints ## Deployment Caddy is already running with the new config on indri. This PR captures the ansible changes. ## Next Steps - Update zk docs with new git remote format - Migrate registry and other services to Caddy - Retire tailscale_services ansible role 🤖 Generated with [Claude Code](https://claude.com/claude-code) Reviewed-on: https://forge.tail8d86e.ts.net/eblume/blumeops/pulls/56
2.9 KiB
2.9 KiB
Tailscale Kubernetes Operator
Manifests for the Tailscale Kubernetes Operator, managed via ArgoCD.
Source
operator.yaml- Static manifest from https://github.com/tailscale/tailscale/tree/main/cmd/k8s-operator/deploy/manifests- Secret block removed from
operator.yaml- managed separately viasecret.yaml.tpl - Image reference changed to fully-qualified
docker.io/tailscale/k8s-operator:stable
Prerequisites
- OAuth client in Tailscale admin console with:
- Devices: Core (Read & Write) - tag:
tag:k8s-operator - Auth Keys: Read & Write
- Services: Write
- Devices: Core (Read & Write) - tag:
- ACL with
tag:k8s-operatorowningtag:k8s(so operator can tag resources it creates)
Manual Bootstrap (Before ArgoCD)
Tailscale operator must be deployed before ArgoCD since ArgoCD uses Tailscale for ingress.
# 1. Create namespace
kubectl create namespace tailscale
# 2. Apply OAuth secret (uses 1Password)
op inject -i argocd/manifests/tailscale-operator/secret.yaml.tpl | kubectl apply -f -
# 3. Apply manifests via kustomize
kubectl apply -k argocd/manifests/tailscale-operator/
Ongoing Management (After ArgoCD)
Once ArgoCD is running, the operator is managed by the tailscale-operator ArgoCD Application.
ArgoCD pulls manifests from forge and applies them automatically.
ArgoCD CLI Commands
# Check application status
argocd app get tailscale-operator
# Trigger a sync (pull latest from forge and apply)
argocd app sync tailscale-operator
# Preview what would change without applying
argocd app diff tailscale-operator
# View deployment history
argocd app history tailscale-operator
# Hard refresh (clear cache and re-fetch from git)
argocd app get tailscale-operator --hard-refresh
Verification
# Check operator pod is running
kubectl get pods -n tailscale
# Check operator logs
kubectl logs -n tailscale -l app.kubernetes.io/name=operator
Files
| File | Description |
|---|---|
kustomization.yaml |
Kustomize configuration for all manifests |
operator.yaml |
Operator deployment, CRDs, RBAC (secret removed) |
proxyclass.yaml |
ProxyClass with fully-qualified images |
dnsconfig.yaml |
DNSConfig for cluster-to-tailnet name resolution |
egress-forge.yaml |
Egress proxy for accessing forge on indri |
secret.yaml.tpl |
1Password template for OAuth credentials (manual) |
README.md |
This file |
Notes
- TODO: The OAuth secret (
operator-oauth) is not managed by ArgoCD and must be applied manually. Future improvement: integrate with a secrets operator (e.g., External Secrets). - Services using the Tailscale LoadBalancer should reference the ProxyClass:
annotations: tailscale.com/proxy-class: "default" - The egress proxy for forge is deprecated. Forge is now accessible via Caddy at
forge.ops.eblu.me(HTTPS) andforge.ops.eblu.me:2222(SSH), which pods can reach directly.