Add Kubernetes Service for Dagger Engine

Context

We operate a large-scale software factory supporting key business units (DSM/DSI).
To enhance our CI capabilities and accelerate delivery, we are deploying Dagger as our new CI solution.

Our setup:

• Dagger runs as a DaemonSet in Kubernetes
• Shared Dagger modules live in GitLab
• Projects use these modules via GitLab CI pipelines

This aims to provide a scalable and reliable CI platform aligned with our business objectives.

Current Limitation

The current Helm chart exposes Dagger only via a NodePort configured directly on the pods (DaemonSet or StatefulSet).
As a result, each CI pipeline must explicitly target a specific node IP, for example:

_EXPERIMENTAL_DAGGER_RUNNER_HOST: "tcp://<node_IP>:<dagger-port>"

This approach:

• Strongly couples the CI to the cluster's topology
• Prevents load distribution across nodes
• Reduces resilience, e.g., if the targeted node is restarted

✅ Solution

This PR introduces:

• A Kubernetes Service to expose the dagger-engine pod :

_EXPERIMENTAL_DAGGER_RUNNER_HOST: "tcp://dagger-engine-service.dagger.svc.cluster.local:<dagger-port>"

• Session affinity (sessionAffinity: ClientIP) to ensure that all requests from a pipeline are routed to the same pod, maintaining execution consistency.

Changes

• Added engine.service.* configuration in values.yaml
• Created new Helm template: engine-service.yaml
• Added app: label to pods so the Service selector can match them
• Modified pod templates to skip hostPort configuration if service.enabled is true

Goal

This change enables a more scalable, resilient, and Kubernetes-native integration of Dagger, and removes the need to target specific nodes from the CI.