CI lacks end-to-end testing with real Kubernetes cluster

[yairfalse]

Problem

Current CI (.github/workflows/ci.yml) only runs:

• Format check
• Clippy
• Unit tests
• Security audit
• Build

There's no validation that the controller actually works in Kubernetes.

Impact

• Regressions in K8s API interactions won't be caught
• Gateway/HTTPRoute reconciliation bugs can ship
• No confidence that releases actually work

Suggested Fix

Add a CI job that:

1. Spins up Kind cluster
2. Installs Gateway API CRDs
3. Deploys RAUTA
4. Creates test Gateway + HTTPRoute
5. Sends HTTP traffic and verifies routing
6. Checks controller logs for errors

Example workflow addition:

e2e:
  runs-on: ubuntu-latest
  steps:
    - uses: actions/checkout@v4
    - uses: helm/kind-action@v1
    - name: Install Gateway API CRDs
      run: kubectl apply -f https://github.com/kubernetes-sigs/gateway-api/releases/download/v1.0.0/standard-install.yaml
    - name: Build and load image
      run: |
        docker build -t rauta:test .
        kind load docker-image rauta:test
    - name: Deploy and test
      run: |
        kubectl apply -f deploy/
        # ... run integration tests

[yairfalse]

Still valid. We now have an oracle (eval/oracle/) that validates the gateway's external behavior against a live instance (21 test cases). K8s-specific e2e testing (kind cluster + Gateway API CRDs) is the next layer. The oracle pattern from AHTI/Rauha can be extended with a K8s Job definition.